Wow. AMD has really pulled through on this one. They are really giving Intel a run for their money. This is going to be great for consumers. What is interesting is the HUGE jump in multi-threaded performance on AMD. Very interesting.
They're coming in at half the price of equivalent Intel chips. They don't need to win, necessarily (which I didn't think was going to happen), they need to be competitive for a lot less money.
EDIT, in response to the other edit: sorry about that, I wasn't sure which comparison you were making. Still, the difference is slim enough that if I were in the market, I'd go with this one just to put the heat on Intel.
You do have to consider that this is an 8 core, and more and more applications are being developed with this in mind. Whoever still thinks that Single threaded performance is the most important, lives in the past.
Sure, but Excel is presumably one of the most optimized (in terms of the amount of money and brainpower devoted to optimizing it) applications in existence.
That's the price of backward compatibility. But I think it's a bit of a "Seinfeld is unfunny"[1] situation; if Microsoft decided to save on the nine-figure budget and outsource Excel development to, say, your favorite purveyor of antivirus software, you'd notice real fast.
That's not true. 2 weak horses != 1 strong horse. 1 strong horse can do the work of 2 weak horses, but not the other way around. A single core can run multiple threads, but multiple cores can't work together on a single-thread task.
I'm definitely going to be building a PC with one of these CPUs some time this year. I've been holding off on an upgrade because I couldn't justify the cost. This is a lot more reasonable.
Quad cores are the absolute minimum I would want for a standard office machine, unless it's only for a handful of basic applications. My latest work laptop is a quad core with 32 GB of RAM for running multiple VM's, but they couldn't give me an SSD which ends up wasting at least a few hours of productivity a week.
Right, I guess that's my point. The minimal gains in CPU are widely offset by upgrading other components like a SSD. So what if Ryzen is a few percent slower on a javascript benchmark? Does this actually noticeably affect any apps that people actually use?
By a pretty small margin compared to equivalent Intel CPUs from what I've seen[1]. And the ryzen chips are a whole lot cheaper and beat Intel for parallel tasks.
I'm about to buy a beefy machine for FPGA development, I'm definitely considering AMD. Seems like I'd get better performance for a much lower price. I'm still waiting for more in-depth reviews though.
It's interesting that you'd mention that workload as it's traditionally been completely single threaded (where it matters), and exceedingly memory intensive. Do you have first hand experience with how well it scales with cores?
Because people still buy new computers and use those applications on them – that's why single-thread performance is (for the moment at least) still important.
I do understand where you're coming from, but real-world performance is important, especially when that world is imperfect.
For these applications, any modern desktop CPU is enough, and enough is all you need for them. People browse the web and do office tasks just fine on tiny 10W "Ultrabook" CPUs.
Just this week I bought three systems for PXE-boot office use at http://infinite-food.com/ and settled on Kaby Lake G3950 dual core Celerons with integrated Intel HD graphics. Here in China a whole mini-ITX system (including case, 180W power supply, 8GB DDR4 2133 RAM, G3950 CPU, motherboard, keyboard and mouse but not screen) is ~USD$174 (inc. parts shipping). I think Intel's integrated GPUs are still very attractive at the lower end of the market.
> Because people still buy new computers and use those applications on them – that's why single-thread performance is (for the moment at least) still important.
That's a load of bullshit. No one ever complained that the web is unbrowsable or an excel spreadsheet is unworkable without a high-end boutique processor. Those benchmarks are simply a load of bullshit that evaluate something that is utterly irrelevant.
Hi geezerjay, I'm trying to find a way to contact you - I was wondering whether you could comment on a previous comment about used opterons for "Budget HPC": Where would I go to find this, and build myself such a rig? Links?
Because enterprise sales dwarf consumer sales? Enterprise is replacing servers and workstations every 3 years. Your 7 year old home beater will run until it dies.
With AMD's previous offerings, they benched well but ran like crap because so many applications are single-threaded. These single thread tests are important to know how these chips will work in real world scenarios.
This is the data enterprise buyers want to hear. If AMD can match Intel at a much lower cost, expect enterprise buyers to switch.
If single-threaded benchmarks didn't matter, then bulldozer would have easily won with its smaller, less efficient cores (it was actually best for a while at some very parallel workloads).
That said, those numbers aren't particularly bad for a new CPU running un-optimized software. All the important software will be updated soon enough once the compilers are in order.
Well yeah but will you buy an alternate CPU that might be faster or might be slower in some games, or will you buy standard Intel, for which every game was definitely tested and will run with standard performance?
It's also 0.1fps faster than the 6950X in the Civ VI AI test. And the 6950X is literally fucking $1700, that's over THREE TIMES more expensive than the top Ryzen.
But you ignore that the i7-7700K is faster in that benchmark. And that one is cheaper. So is the i5.
If you do more than games, the Ryzen seems like a a good choice. It has serious power in applications, and is still fast in Games - which is a big step up compared to the FX. But no, it is not the goto choice for a gaming PC.
Yeah, I'm not arguing with that, I actually made that same exact point in a different thread :) I'm just pointing out how it compares to more expensive 8/10-core CPUs.
If the 6C/8T and 4C/8T parts will launch at the same frequency as the high end (and i've read that at least for the 6C they will) we'll end up with a CPU that is as fast as the 1800X in games (which is good enough no matter what the 120 fps enthusiasts say) but this time is cheaper than the 7700K. That will be the sweet spot for just gaming performance, not these ones.
I'd take Intel CPU that runs 120 FPS average over AMD that runs 130 FPS average, when lowest FPS for Intel drops to 100 FPS and for AMD it drops to 80 FPS...
> But it's still losing in every web and office single threaded benchmark.
By ~15% compared to >160% for Bulldozer.
And this is day zero before any applications have been optimized for the new microarchitecture, compared with Core which everything has been targeting for more than a decade.
> Would you do it if you had the choice between that or adding new money making features/fixing bugs?
Of course I would, because people are going to benchmark my software against my competitor's software and choose the one that runs faster on their hardware.
Especially if my software is a compiler and people are benchmarking the programs it produces against other compilers, which is where most of the optimizing actually happens these days.
Give compilers some time and the optimizations will be "free" to developers.
Besides, if your application is one where performance is a serious issue, then you are going to spend some time optimizing rather than adding features. For some applications, performance is a feature.
People always seem to forget this in terms of CPU optimizations. It's both good and bad, though. If you're pegged to a given version of a compiler, the point releases likely won't see any optimization while the major releases will likely include additional optimizations for newer chipsets. And I'd wager AMD will be working to get optimizations into both open source compilers as well as things like Visual Studio. One thing to note is that switching to a newer compilker will likely be easier for most of Linux-land as opposed to Windows land. Upping the compiler version with your code compiled in GCC is often much easier than switching to a newer version of Visual Studio due to both cost and dropping older operating systems (the latest won't support Windows XP and Vista for instance, which may matter to some specific business lines and consumer areas).
But it wins in every case where performance truly matters: multi-threaded calculations, archiving, encoding. You know, the things that actually do take up time.
Do you really care if your webpage renders for half a second longer?
No, it doesn't. Their hyperthreading model is clearly not as good as Intel's. Take a look at the Handbrake tests, which take full advantage of hyperthreading. The 7700K, with half the cores, is only a few frames/sec. behind.
If I was after encoding performance, I wouldn't buy a Ryzen. If I was after rendering performance, I wouldn't buy a Ryzen. Throw an mid-range nVidia on there for OpenCL, and the only CPU that makes sense for that price range is the 7700K.
Looking at most benchmarks hyperthreading model is equal or better on Ryzen.
Ryzen 1800x and 6900k have almost the exact single threaded score on Cinebench. But 1800x beats 6900k on Cinebench and Handbrake
The reason he said that is because it makes infinitely more sense to use a renderer that does it's work on the GPU with CUDA or OpenCL.
Only if you are using a shitty software renderer or are completely stupid would you do rendering on the CPU.
The Ryzen performs nicely on the encoding tests, though that will be performed by the GPU soon. It already can be (NVENC and AMD VCE), but the quality isn't quite as good as everyone's favorite h.264 encoder, x264.
Depending on rendering task, GPU rendering isn't always feasible. In particular, for very large scenes consuming a large amount of RAM, it is difficult to find graphics cards with over 8 GB of VRAM, while 32gb or 64gb of desktop RAM are easily accessible.
Yes, I do care. A lot. Actually if it's just 50ms I care. Slow main thread is a huge annoyance, and that's 99% of my computer use. On the other hand, I don't care if asynchronous tasks (encoding etc.) take 30 or 40 seconds or whatever; I can get a cup of coffee or check my email. Also, major parallel workloads are done on the GPU anyway.
What? I've never run a parallel op on the GPU but I do on my CPU multiple times a day. GPU acceleration besides games is still extremely rare. You're grasping straws here.
You're also missing that even on chrome the js and layout engines run different threads. Same with audio and video rendering. Same with downloads. The average page is probably running things on 10-20 random threads. And the average user has multiple tabs open.
Multicore CPU is so common these days that any performance critical code that can be parallelized, is. This includes GZIP and HTTP which are used heavily by web browsers. The network stacks underneath those in the OS are also multi core capable.
Single thread performance is not all that important anymore. It's better to have 2x the number of fast cores than half the number of slightly faster cores in almost every situation these days.
Generally speaking the price of a processor is going to be the square of the single threaded performance times the number of cores. So your choice is more like one 8 GHz core versus 4 4GHz cores. Which might be good or bad depending on your use case. Well, there's a threshold below which you only get linear per-core scaling but outside microcontrollers you're not going to have to worry about that.
The R7 1700 is within shouting distance of the 6900k in most serious tasks - rendering, encoding, compilation etc. A $330 65w CPU is a credible alternative to an $1100 140w CPU. That is absolutely remarkable. The cherry on the cake is that the cheapest AM4 motherboards are half the price of the cheapest 2011-v3 boards.
For the first time in a decade, AMD are seriously competing at the high end. That's great news for everyone except Intel.
The Zen based APU platform is going to cause real problems for Intel in the laptop/ultrabook market, Ryzen's power consumption looks very low and their iGPU technology is years ahead of Intel's.
>> The Zen based APU platform is going to cause real problems for Intel in the laptop/ultrabook market
Technically this is true, but unlike the desktop market where you can buy stuff in pieces, it really depends on how the laptop manufacturers adopt them. As in how fast they show up in product lines, and in which laptop segments they show up in.
If, for example, Intel gives laptop manufacturers pricing incentives (legally or not) to stick with their chips, it may not cause any problems at all with respect to actual customer purchases.
I'll have to disagree on a few of the points on that list:
> - relative multi-threaded performance per $
Highly depends on the particular workload. Given the few benchmarks we have right now, it seems to go 50-50 once you factor in the lower price of the 7700k. At the very least, this one is close.
> - memory bandwidth
Seems about equal from the measurements I've seen (the theoretical numbers are irrelevant).
> - L3-cache bound workloads (AMD has twice more L3)
Cache-size bound workloads to be precise. Cache latency still seems higher, even on L3.
Overall, the 1800X seems to make the entire Broadwell-E series redundant on price/performance (at least at current price), no argument there. Against the 7700k, however, it's a very interesting decision, entirely based on the distribution of your workloads. And if Intel drops the prices on Kaby Lake, the value equation can change very quickly indeed.
Guru3D seem to be the only ones to manage memory frequencies over 3 Ghz, most reviewers report a lot of issues with the memory subsystem at the moment. In many reviews (see [1] for a representative example) the 1800X is on par or even slightly behind the 7700k in memory bandwidth, so at the very least the situation is far from unambiguous.
And? Intel's $1,600 6950X loses to their own $170 i3-7350K in single-threaded benchmarks. When you pay extra for a CPU with more cores, you almost always get less single-threaded compute because of power/heat/design constraints forcing the clock speed down. Dual cores will often beat quad-cores, which will almost always beat hex/octo-cores.
What's happening here is that AMD has priced Ryzen so inexpensively that it actually broke the usual market segmentation by driving high-thread-count workstation CPUs (usually $1k+) into the same price range as high-end gaming CPUs (~$300-400). Think of it as being able to buy an industrial dump truck for the same price as a pickup truck. The dump truck is clearly a better value in terms of hauling capacity, but most people would be far better off with a pickup truck for their daily traveling.
AMD's pickup truck (the 1400X) doesn't launch until later this year. The 7700K will probably still beat it on single-threaded workloads due to the difference in clock speed, but it should get you most of the way there for substantially less money, around $200 if the leaks are accurate.
I think Intel severely underclocks the 6950X by default, so that it fits within the stated TDP of 140W with all 10 cores at full load. If you have adequate cooling, it is very easy to overclock. I think I ended up at 4.2GHz for 10 core loads, 4.4GHz for 4 cores.
I think people run enough browser tabs (+ their office suite + background OS tasks) where the core difference will outshine the single-core speed. It's not hard to overwhelm 4 real cores.
Ryzen 7 chips focus on heavily multithreaded work. When you are dealing with every day apps and games you are more concerned with IPC (instructions per cycle) and I'm sure we'll see better performance in that arena with Ryzen 3 and 5
> What is interesting is the HUGE jump in multi-threaded performance on AMD. Very interesting.
Indeed. Single threaded performance, however, seems bad. Virtual cores most likely share less of the actual core than on Intel parts, allowing one to compete less with the other and get more done. On my i7 laptop, on anything more demanding, I see half the "cores" idling, probably because one is using the actual underlying core while the other is waiting.
Still, while AMD obviously did a good job, it will not make the Intel guys lose any sleep for now.
The server parts are another story. If they manage to reliably outperform Xeons (even if it's only on a per watt basis), it'll be fun to watch.
Hyperthreading is not a virtual core and as it doesn't exist as a separate core its hard for it to be idling. Bad analogies make for less than adequate understanding.
They are also called "logical cores" (I think Intel documents prefers "logical" and Sun - they invented the concept - prefers "virtual"). One logical core can appear to be idling to the OS if the parts of the physical core are in use by another (SPARC and POWER go up to 8) logical core (as it's resource starved). It seems, from the benchmarks, the AMD parts have less shared resources between logical cores (or a finer granularity for allocating) than the i7 in my laptop.
Pre Ryzen AMD didn't have anything like hyperthreading. The prior generation had n physical cores wherein each 2 shared a fairly significant chunk of hardware. The new generation of AMD and many generations of intel have extra hardware in each physical core that makes it look and act like additional cores. If I understand it correctly, and please correct me if I'm wrong, it mainly serves to ensure that the core is better utilized in general.
From the benchmarks, it seems a single thread can fully utilize all resources from an Intel processor while it can't on the new AMD ones and that's why their single-thread performance is lower and their multi-thread performance is higher.
We should keep in mind the benchmarks know nothing about the new AMD parts and how to keep all parts of the processor busy.
> Many of the cases where I care about perf are cases in which the app is multithread savvy
Yeah, as would be expected from most people on this site. Sad thing is, there's a bunch of pissed-off gamers right now who were on the hype train expecting Ryzen to be the next big thing in gaming, but it's far from it. Ryzen is great, just not amazing for gaming.
Still Waiting for a real person that has a 'real chip' that has tested this thing and bench marked it for 'real'.
In so many years, I have never seen so much marketing where nothing actually exists in peoples hands, well since the introduction of 'segway', that took months of hype before we found out what it really was.
If it be true that this chip is what they say, then when will we see real benchmarks by real people? They're saying "March 5", yet the marketing stuff makes it appears that AMD already won, hell they're not even at the starting gate, hey I'm rooting for AMD, I wish this all to be true, but I'm not spending +$1,000 bucks for a new MB & Ryzen, until I know for a fact this thing is real.
Impressive! While it doesn’t utterly destroy Intel, AMD does offer a MUCH better price/performance ratio. Things have gotten much more interesting indeed. Intel’s dominance is being called into question. As a result, we all profit.
On the other end ARM is also nibbling at Intel with AARCH64 chips that are closing on lower-end Core performance and building in server and high-end workstation grade features like virtualization.
Yeah, and AMD is the only company who can offer a high performance GPU together with a high performance x86 CPU. And APU's with HBM of course. Looking forward to what they can achieve now that they seem to be back (although most of the financial success probably lies with the server CPU Naples).
One can hope that in the long run we'll see more things with a SPRI-V back end and more libraries in that space to put AMD and NVidia GPUs on more equal footing.
Something like that could be a monster for HPC & ML type tasks. Provided AMD gets their software and tooling up to par; Nvidia doesn't rule HPC/ML because their GPU's are leaps and bounds ahead of AMD, but because CUDA is..
If so, might explain why they went with a (relatively) gimped vector FPU for Zen compared to Intel offerings (particularly the soon arriving Skylake Xeons with 512-bit wide SIMD).
It offers much better price/performance ratio if the seriously multi-threaded use scenarios matter for you (that is where you compete with Intel's seriously overpriced >$1000 chips).
It seems that in common use (web browsing, office, gaming), fewer but stronger cores still shine and even the fastest Ryzen is slower and more expensive than Intel's offering.
Well...for the time being. When the R5 and R3 get released in Q2, we will have 6 and 4 core versions with higher clock speeds so that might help a bit. Plus based on the pricing, I'm gonna bet a 4 core ryzen will be a bit cheaper than a 4 core kaby lake.
No it doesn't. Chips much slower than Ryzen aren't struggling with those common usecases at all. All you're doing is losing 8C/16T and spending the same amount or more to do it.
> ... this means that the base memory controller in the silicon should be able to support ECC. We know that it is disabled for the consumer parts, but nothing has been said regarding the Pro parts.
ECC should be a standard feature. If you don't want it you don't have to use it but to disable simply for market segmentation is lame.
Is that buffered ECC or unbuffured ECC? Because last time I checked, my AM3 board had support only for unbuffured, which is really hard to find and I just didn't bother.
I don't know why you're being downvoted, your comment is accurate. Consumer hardware traditionally does not support registered memory, which is (reg ECC) the standard for server hardware hence cheap, while unbuf, unreg ECC is rather rare. AM4/Zen is no exception, no DDR4R for consumers.
FWIW, I remember ASRock being mistaken about AMD features in the past: iirc there was a while when they were wrongly asserting that Bulldozer didn't support I/O virtualisation. (Eventually they implemented it on their own AM3 motherboards.)
Thanks for pointing this out. I found this on the specification page[0] for the ASRock "X370 Killer SLI/ac" AM4 board:
> AMD Ryzen series CPUs support DDR4 2667/2400/2133 ECC & non-ECC, un-buffered memory
So the board seems to accept ECC memory, hopefully that means the memory controller actually performs ECC? I found a reddit comment claiming that the boards accept ECC memory but don't perform error correction[1]. Not a great source but seemingly possible.
Do you have any more sources suggesting that ECC is actually being implemented?
For example, BIOSs often have ECC settings, if we had some screenshots of an AM4 board bios showing ECC setting that would be strong evidence.
That's good information but still doesn't tell us that these cpus support ECC, right? Maybe the motherboard fully supports ECC when used with a future 1800X-PRO but not with the 3 chips reviewed here.
In that case anandtech saying "We know that it is disabled for the consumer parts" and ASRock saying "ASRock AM4 motherboards fully support ECC function" can both be true.
AMD will use the same die for its server CPUs, so ECC support is on chip anyway. I see no reason why they should turn it off in the Ryzen CPUs launched today.
The IMC in almost any Intel CPU also could do ECC, but on most consumer SKUs there's that pesky little configuration flag set by the factory telling it not to. Why? Because Intel can, that's why.
Gigabyte mobos say you can use ecc memory but not run them in ecc mode.
"Support for ECC Un-buffered DIMM 1Rx8/2Rx8 memory modules (operate in non-ECC mode)"
> So the board seems to accept ECC memory, hopefully that means the memory controller actually performs ECC? I found a reddit comment claiming that the boards accept ECC memory but don't perform error correction[1].
This is set up by the firmware (BIOS). Since we all already know that quite some of these still have issues (unsurprisingly, it's a new platform), it isn't far-fetched at all that current firmware doesn't quite do the right thing with certain memory sticks.
Note that this is unbuffered, unregistered ECC memory, which is usually more expensive and less available than the standard server memory (DDR3R / DDR4R), which is NOT compatible with any consumer CPUs.
IIRC just because it boots with ECC ram doesn't mean it is actually error correcting. I'm under the impression Intel spends considerable amount of money to prove their memory controllers actually error correct & historically AMD hasn't spent the money on documenting theirs.
Yes, I have not been able to find any data from AMD directly that ECC is not or is supported. In any case I am disappointed. Desktops, workstation, servers are all getting more and more memory. The changes for corruption are only going up. And getting an Opteron for each workstation is kind of silly.
Edit: If no motherboard supports its, in Linux one can use ecc_enable_override (similar thing might be available in Windows). That is if the CPU support it. It can USE ECC memory, but we don't know if it can use the ECC checksums, most motherboards have a NO here.
Looks like you figured it out. If not, I did actually see someone about a month ago mention in a comment here, that they may or may not be in possession of a Reddit account. I'll take a look and see if I can find the original comment. Maybe someone else has seen.
What means validated in this case? It's not certified by AMD to work, so depending on the memory and motherboard combination it could work, or if you are unlucky it doesn't work?
Validated means run it through server/workstation grade testing. For the first Ryzen processors, focused on the prosumer / gaming market, this feature is enabled and working but not validated by AMD. You should not have issues creating a whitebox homelab or NAS with ECC memory enabled.
Thanks. Still want to see someone actually get ECC to show up as enabled on something like the Asus Crosshair Hero board, which is what I've pre-ordered. I just grabbed a 16GBx2 kit of G.Skill 3200 but it would be great to get ECC for this quantity.
>ECC should be a standard feature. If you don't want it you don't have to use it but to disable simply for market segmentation is lame.
I agree ECC should be a standard feature, but it follows the same marketing patterns (in that industry) as disabling multi-processor support or artificially limiting clock speeds.
I think ECC is in a different category than MHz or core count because with a lower frequency or fewer cores you may get your result slower but it will be the same result as a top end chip.
Without ECC you may get a completely wrong answer when the increasingly small, vulnerable, and numerous dram cell is flipped.
Actually no, enabling external cache coherency (eg. QPI) incurs a significant cost in terms of pin count, power, die space and clock speeds. The same does not transfer to ECC memory.
This is done so that whitebox system builders don't buy consumer level motherboards and CPUs, shove ECC in there, a cheap raid controller, and sell this setup as servers. Next thing you know all the AMD servers on the market are failing because of shoddy parts because consumer level QA is at a lower level than enterprise, not to mentioned the lowered performance of consumer level hardware. Now AMD has a bad reputation for reasons completely unrelated to its chip quality or performance and collapses in the market.
I don't like this situation, but I certainly understand it. AMD needs the server market to raise revenues. This means its server offerings, from soup to nuts, need to be rock solid. Letting whitebox resellers junk up their reputation is not within their interests.
Not to mention, enterprise sales often subsidize consumer sales. Those big enterprise markups go into the same pot of money that allows AMD to sell consumer chips at lower prices. Anything that threatens enterprise sales could affect the consumer space. AMD needs to be very careful here as, compared to intel, it needs to prove the value of its brand every chip generation. They also need to overcome a lot of bad mojo they've developed from questionable benchmarking and marketing of its previous two, or more, generations of chips. Bulldozer, I believe, had half the single core performance compared to the intel equivalant at the time, yet it benchmarked well due to its extra cores. The market doesn't forget shenanigans like these.
That said, the market allows 'workstation' level cpus and motherboards that support ECC. A little legwork can find you decent deals, but no, they're not going to be $59 NewEgg specials and you won't be able to use the consumer Ryzen on it. Consumer CPUs are almost always gimped anyway - smaller caches, disabled features, slower clocks, etc. The ECC limitation isn't too different from this, but it is annoying.
The 1700 is really well placed, often seems to beat the i7 7700k in benchmarks whilst drawing less power and costing roughly the same/cheaper. Probably even better value if the stock coolers are as solid as they sound. Definitely going to be a popular chip with gamers (I'm very tempted myself).
Hopefully more software starts to make use of multiple cores more effectively. It's not like it's a new feature of chips any more.
I haven't found one yet, but as the chip is rated at 65W and the i7 7700k at 91W, it's pretty likely to draw less (at load anyway). I'd be curious to see what they draw whilst idle though.
After having a look at some of the links posted, it looks like the power draw is slightly higher than the i7 7700k in most tests, although lower in some. Looks like my assumption based on the TDP was a bit off (also my misapprehension that TDP is power draw, whoops) :)
Especially in games i think it's pretty clear that even the 1800X is not amazing, really good, but behind a 7700k while being more expensive. If you are not solely a gamer though, it's the better choice!
According to AMD CEO Lisa Su at https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea... : "there are some games that are using code optimized for our competitor... we are confident that we can work through these issues with the game developers who are actively engaging with our engineering teams."
Also, another AMD employee chimed in:
"But what's also clear is that there's a distribution of games that run well, and a distribution of games that run poorly. Call it a "bell curve" if you will. It's unfortunate that the outliers are some notable titles, but many of these game devs (e.g. Oxide, Sega, Bethesda) have already said there's significant improvement that can be gleaned. We have proven the Zen performance and IPC. Many reviewers today proved that, at 1080p in games. There is no architectural reason why the remaining titles should be performing as they are." Source: https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
Unfortunately you can bet Intel is going to put all the spin they can into "AMD underperforms for gaming", while the software game developers are going to slowly optimize for Zen, up to the point where this perf discrepancy will be fully fixed but people will still have the mindset that "AMD unperforms for gaming"... sigh
Exactly the opposite is correct. ICC generated code whose runtime dispatch looked at "Are we running on Intel? Select fast path. Not running Intel? Select generic i386 path."
Intel provides a lot of tools and SDKs for game developers. I presume some/many are specifically optimized and highly-tuned for running best on Intel processors:
> while the software game developers are going to slowly optimize for Zen
Honestly I'm not very optimistic about that happening. Maybe some will. But the mentality these days seems to be that there's one winner and fuck all the rest -- we're busy and we don't have time to optimize for losers. Like with web developers seeming to almost exclusively work with Chrome except where they're told to support IE (or Safari?). Firefox just isn't popular.
Still, there's a chance the multi-core revolution will eventually happen. Vulkan could help us get there, although weren't DX12 was supposed to provide similar improvements in multi-threaded access to the GPU?
Also, super important: another reason games underperform is that Windows load-balances threads across CCXs (CPU complexes). The 8 cores of Ryzen are split across 2 CCXs. The 8MB of L3 cache is also split with 4MB per CCX. So when the OS scheduler arbitrarily decides to move a thread from a core to another, it might accidentally resume the thread on the other CCX, causing the thread to lose all is L3-cached data, incurring a high perf penalty.
Boot speed in my experience is mostly i/o bound so I doubt it makes a significant difference over a similar Intel chip. Is boot performance still an issue nowadays anyway? Since I've switched to SSDs all my computers boot up in a few seconds (Linux or Windows). And if you hibernate instead of shutting down it's even faster.
Compiler tests sound more interesting to me, I'm looking forward to those benchmarks.
Boot speed is not even IO-bound for me, it's crappy-firmware-bound. `systemd-analyze plot` shows a total boot time of ~15 seconds, of which ~10 seconds is spent on the firmware, ~3 seconds on the kernel and hardware initialization, and 1-2 seconds on the entire userland including desktop session.
Regarding boot speed, here is quote from Computerbase:
Boards initialization takes forever
For the last weeks Computerbase had access to different mainboards [...]. They had an extremely long init process in common: even with the latest BIOS it took up to 30s to get to the POST screen. Obviously this needs optimization.
-----
German:
Die Boards initialisieren ewig
ComputerBase konnte die letzte Woche auf drei verschiedene Mainboards zugreifen, neben den X370-Varianten von Asus und MSI auch auf das Gigabyte GA-AB350-Gaming 3 mit Chipsatz B350. Allen gemein war auch mit dem letzten BIOS der extrem langwierige Initialisierungsprozess: Bis der Post-Screen erscheint, vergeht über eine halbe Minute. Auch hier gibt es offensichtlich noch deutlichen Optimierungsbedarf.
But the actual compilers people use (IE not that version of GCC) are rarely well-optimized for CPU.
In fact, most compiler profiles are not flat yet, and where your compilation takes a long time, it's because some optimization or algorithm has gone nuts and needs to be fixed.
Until that kind of state of the world changes, measuring compile time is IMHO, a bit pointless, because they are only artificially CPU bound, and not in a way that is useful to benchmark
(IE while i understand that people use these compilers every day, and care about performance, i'm just saying that, given a compiler benchmark, i could still probably make it take zero time with limited amounts of work)
GCC is a useful benchmark not because it is unoptimized, but because it represents a real-world work profile that is otherwise poorly represented in SPEC. The working set sizes are large relative to the cache size, and its composed of a bunch of pointer-chasing branch-heavy logic as it winds through the various trees and intermediate languages. That profile is quite a bit closer to the profile of common business and server applications than nearly all of the other members in SPEC.
"GCC is a useful benchmark not because it is unoptimized, but because it represents a real-world work profile that is otherwise poorly represented in SPEC."
Sure, in that sense, but you also don't want benchmarks that are trivial to game by performing optimizations to them.
:)
For SPEC, you can't edit the source code of the unit under test. Only the compiler and host hardware count. Some components of SPEC have been "broken" by compilers in the past, by implementing some optimization or other that suddenly made that particular test code much faster. But you don't ever get to re-write the benchmark itself, you can't replace a 2^N path with an N^3 path or anything like that. So these kinds of breakages have almost always been restricted to benchmarks that are reducible to one or two relatively tight inner arithmetic-heavy loop(s).
If someone managed to make an optimization that could suddenly make GCC 2x faster without humans editing GCC's source, that would be a major CS breakthrough :)
"If someone managed to make an optimization that could suddenly make GCC 2x faster without humans editing GCC's source, that would be a major CS breakthrough :)
"
No, it really wouldn't.
For example, if you optimize the line-ending and tokenization finding in GCC's parser to use SIMD, you would become significantly faster.
Not 2x, but significantly. These are entirely doable by compilers.
" But you don't ever get to re-write the benchmark itself, you can't replace a 2^N path with an N^3 path or anything like that."
Except, you can, by hoisting calls out of loops, etc.
There are also dynamic memoization optimziations that can be applied, etc.
"If someone managed to make an optimization "
This is a common fallacy that it's ever really one optimization. It's usually combinations of optimizations, each buying you N%.
I will state outright, having worked on pretty much all of the slow parts of gcc (according to spec profiles), and having a team of 80+ people working on compilers, 2x on gcc through compiler optimization is more than within the realm of possibility.
>In fact, most compiler profiles are not flat yet, and where your compilation takes a long time, it's because some optimization or algorithm has gone nuts and needs to be fixed.
For some reason one of my sideprojects needs more than two minutes to link but the task manager tells me it's neither CPU nor RAM nor disk bound. I haven't found the cause yet.
I'm actually curious which prices have changed. A quick google turned up an article claiming prices were being cut [1] but looking at the full price history tells a different story:
It looks like Intel's retail channel partners have a lot of inventory still in the pipeline. Prices are still within the bounds of the second half 2016. My best guess is the retail channel partners will be the most hurt by any slowdown of Intel CPU sales.
And it's mainly because games aren't optimized to take advantage of 16 threads - yet. That's partially Intel's fault for keeping 8+ core processors so expensive.
AMD is trying to change that. Of course 16t games won't be available from day one of Ryzen launch.
As a lot of other people online are pointing out, this is pretty much the exact same thing people said when Bulldozer came out, but it never materialized. AMD has always won in terms of core count, and people have been saying for years that games capable of taking advantage of 8+ threads/cores are right around the corner, but it never really happened. We shouldn't be basing purchase decisions today on what might theoretically happen at some point in the future.
It's much more true today to be fair, and with DX12/Vulkan it should become much more common over the next few years.
Ryzen is much better positioned anyway because it has almost comparable IPC (Broadwell levels), that means even when you get a dual core/single core optimised game you'll get good performance rather than the horrible situation with older AMD CPUs.
They are here now. Almost no modern AAA game recommends less than 4 cores. Many won't start with less than 4 threads. And they do scale to 8 cores. This is because of the consoles, which both have 6-7 cores.
According to the Steam hardware survey, only about half of PC gamers have four cores, with the other half having two cores. Granted, many of those two-core systems will be four-thread systems, but I don't think its fair to say that the market is dominated by 4+ core users today.
All the benchmarks I just glanced over show a different story. Yes, big games claim to want 4 cores because that's what big intel has offered for big gamers. Sure they run on 8-core chips and may get a speed boost for that, or not. But it is quite small overall, with the Intel 4-core parts mostly performing better than Ryzen while being cheaper or similarly priced as Ryzen.
What some people here are anticipating is the situation where games really do scale well to 8 cores, at which point we'd likely see the Ryzen compete against Intel's stupidly expensive 8-core parts and roughly match them in performance while solidly beating them in price/performance ratio. Just like in the encoding benchmarks that actually do run well with multiple cores. These games aren't here now. We aren't there now. But it might be the future.
Don't hold your breath: efficiency at the game's minimum spec is always much more important than efficiency at the top end. And that minimum spec will stay a dual core mobile CPU for the foreseeable future. Therefore, anything that would sacrifice a little performance on a low end few cores system to get even faster on a high end many cores system just isn't an option.
Since I'm not much of a gamer I'm looking at these chips. I wonder what an optimized (packages compiled with flags specific to the chip) Linux system around ryzen would be like.
Would be nice to make a simple one-click benchmark: download a VirtualBox VM and click Start — it boots mfsbsd and starts compiling to a RAM disk, measuring the time.
Too complicated. I was thinking just have some baseline numbers on a clean SSD and run the buildworld a bunch of times. Either would work tho the latter I'm confident I could do.
I'm going to build an 1800X-based box in a few weeks (for games; planning to run Steam on a GNU/Linux distro), I'll want to see `make -j16 buildworld` for sure.
If I read the specs correctly, Ryzens don't have GPUs integrated. Are there any motherboards out there with GPUs on board? Or do you always have to add a 150EUR+ graphics card to it (assuming I want at least 2x 4K display support)?
The 1800X does seem like a killer value for the money for us C++ people :)
Low-end GPUs are a very poor value proposition. For $89 you could have an RX 460 with nearly 3x the performance of the R7 250 plus HDMI 2.0b and DisplayPort 1.4.
The R7 250 is considerably less than enough if you have a 4k monitor. GPUs aren't just for gaming - if you use Photoshop or Premiere, you'll want a decent GPU. An increasing number of applications are making use of GPGPU processing.
The fact that it doesn't have an integrated gpu is a GREAT THING. This is an enthusiast/server CPU, if you can't be bothered to buy a cheap dedicated graphics card, wait for the mobile/apu Zen.
It's mostly also a packaging issue - I don't need 3D capabilities, so a basic 4K supporting Intel GPU works well and doesn't require me to add a GPU to a chasis (which can then be mini-ITX).
Hence my question about having GPU integrated to the motherboard like older boards had it.
I think it'd be great if we could rely on every part having a standard, reasonable, GPU, even if we just use it as a clever vector unit and never push a single pixel through it.
No really. You'd prefer to use that silicon area to increase your perf. If you want a CPU-GPU combo wait for the laptop/mobile part of Zen codename RavenRidge.
There were rumors of slightly better single threaded performance. Stock movements are always relative to expectations, investors knew roughly what was coming but not the exact details.
The idea behind this, simply put, is that if you are in a trade for whatever reason with a stock that has a major announcement coming; you are pretty much at the potential peak at announcement time so it is a good time to get out.
Maybe knowing consumers will probably have more than 2 cores will make games optimise better for parellelism, it's not like the engine programmers _cant_ do it, they do it all the time with GPUs.
GPU parallelism is very different from main-core parallelism. The GPU is an entirely different architecture than the CPU. The stuff that's still on the CPU are things that don't trivially parallelize to begin with.
When making AI bots it's quite common to have all of the work of processing the AI it's own thread, so each rendered frame will do all calculations for all bots. That in of itself would be cheap, except it has to share it's core with a bunch of other crap.
With 2 cores you can have a main thread for updating the screen and then an "everything else" thread. If you can't be sure people will be playing with >2 cores then you have to keep your budgets down to 2 cores when developing and that's quite limiting.
it's not that it's "hard" to parralellise(?), it's that you run out of resources super quick if you make assumptions about CPU topology.
A Broadwell with 3300 Mhz (3700 Mhz boost) beats a Kaby Lake (Skylake...) with 4200 Mhz (4500 Mhz boost)? Can the Broadwell use the 128MB EDRRAM from the iGPU as a L4 cache if the iGPU is not used? Would that actually make such a difference?
> Can the Broadwell use the 128MB EDRRAM from the iGPU as a L4 cache if the iGPU is not used? Would that actually make such a difference?
That's exactly what it does, and it gives you a "free"* ~20% increase in memory operations performance (which tends to be a ~20% increase in general performance).
* not actually free because you paid for those 128MB EDRRAM...
Nice, now I'm wondering why Intel doesn't slap 128MB EDRAM on an Kaby Lake i7 4 core 8 thread and then release it for the enthusiast platform x99. Maybe even select chips to make it possible to release them at 5ghz. Of course with a large premium (expensive EDRAM etc.).
Because they had no competition, so no desire for higher performance. They can easily remove the GPU from Kaby Lake, use that transistor/thermal budget for 4 additional cores and add 128MB of L4 cache for a monster CPU.
The Iris chip (the extra EDRAM for GPU/L4) is seen by Intel as a performance improvement for mobile solutions (ultrabooks, MS Surface Pro, maybe Macbooks some day?) and not a desktop solution.
But from a consumer point of view, it increases performance (either for graphics or for the whole system if I have a discrete GPU), and it can also help with GPU numerical processing, because you get "free" transfers from memory (L4) to graphics memory by flipping a bit (or a few bits) in hardware. Don't know why this wasn't expanded further, appart from the cost of the embedded RAM.
This is what I got frustrated with and decided to stop waiting & buy a Ryzen, in fact. I've been looking forward to an enthusiast segment desktop chip with eDRAM and Iris/Iris Pro (or better!) from Intel, for a few years now, but it hasn't surfaced.
It's hard to put any faith in the large spread of numbers presented by this article regarding gaming performance. Anandtech just recently did a review of the I7 7700K and pretty much all the cpus over $200 posted the same numbers since most games are gpu bound.
Does anyone know if the Zen 1800 desktops support their new Secure Encrypted Virtualization (SEV) extensions or are they only for server class machines (Naples)?
I'm really interested in these from a virtualization standpoint, especially with 8 cores for so little $$. I'd love to see a good set of tests/benchmarks with a completely virtualized workload.
Speaking of which, Zen/Ryzen was supposed to come with Secure Encrypted Virtualization (SEV), but I don't know if any of these reviews have covered it yet:
I'll definitely consider 1800X for my movie rendering/machine learning rig, especially if they confirm ECC functionality. 40 PCI-E lanes would be better, but I can live with 2x x8 GPUs and only one x4 SSD for machine learning.
Interesting would be 4C/8T parts - if they can run them cool (given 8C/16T 1700 uses 65W I think it's not unreasonable), they probably can clock them to get to 7700k parity at 95W.
No, but if they have thermal and power headroom, the X chips will add 100MHz above the "boost" clock. They call it XFR (eXtended Frequency Range). So having a water cooler will make it more likely to get that last 100MHz, but I don't think it's going to make a big difference if you're not overclocking.
What is the deal with 16 PCIe lanes? With the move to PCIe storage, this is disappointing. I am happy to see the price/performance is good (which is surely going to force Intel to be competitive again), but if this is targeting the high-end, I can not understand 16 PCIe lanes.
The CPU has 24 lanes and the chipset can add up to 8 additional lanes, see [1] first table.
It's a pity that they did not include at least 32 lanes on-chip to allow two full x16 GPUs at least on the 1800X, but I hope there is a 1900X in the pipeline with more lanes.
Really not wanting to poop on anyone's parade but an off lease Dell/HP/Lenovo workstation used from eBay will beat the pants off anything else for homelab. You can get an E3-1240 V2 with 8Gb RAM for the kingly sum of ... $200. Including motherboard, chassis and PSU.
And you're also stuck with a $200/month power bill and the full home heater....
Don't get me wrong, I have a bunch of rack pull servers and they're great for the price but I colo mine for less/server than the power bill/server would be at home and have them ona faster connection. I stick with low power options for light in-home usage
The single-socket Xeon E3 that GP mentioned is 69W TDP. The Ryzen 1700 is 65W, while the 1700x and 1800x are 95W. Why do you think the Ryzen box will be a lot less power hungry?
The Ryzen 1700 has 8/16 and much better performance compared to the mentioned Xeon E3. I think the Ryzen 5's would be the ones to look out for (still more cores and much less price), hope they push the power consumption even lower.
It doesn't matter what the TDP is. Home systems like HTPCs and home servers will run at their full power a very small amount of the time they're running. That means their maximum power usage (which is what TDP is supposed to indicate) does not matter.
Modern processors do some clever tricks to minimize using power if it is not needed, and have gotten progressively better at it. For home use cases, it's often worthwile to spend more money to get newer or otherwise more efficient PC components. This is especially true in areas where electricity costs are higher (e.g. average price per KWh can be almost twice as high in the Benelux when compared to the US).
One thing to get an indication about how good a CPU/chipset is is idle power usage of a system. It seems Ryzen is even better at this than Intel's already excellent Sky/Kaby Lake architectures [1], possibly because of the fact that Ryzen motherboards contain fewer components. On the contrary, the Ivy Bridge architecture of the chip mentioned by the parent cannot be considered competitive from a power consumption perspective, despite still being a great performer.
"beat the pants"? Compared to Ryzen, E3-1240 V2 has: half the cores, half the L3 cache, half the memory throughput of Ryzen, no AVX2, etc. Plus average price is more like $300-350 http://www.ebay.com/itm/Dell-PowerEdge-R210-II-1U-Rackmount-... And it's still comparing apples to oranges favoring Intel. Wait a few months and compare used Ryzen gear to that old Intel gear...
Depends on your needs, dual socket xeon servers can be had with huge amounts of ram if you shop around. E.g. I got an IBM 3650 m2 with 2x L5640 and 128gb of ecc ram for $200. For my needs (heavy data processing on a budget) its been awesome.
For anyone looking at a new AM4 build with high speed memory, be warned that Linus' review stated he could not POST with memory clocked past 2666MHz[0]
Is this uncommon? I have a 6950X with some "3200MHz" RAM, and it definitely doesn't work reliably at that speed. Nothing above 2666MHz will pass stress tests, even with conservative timings.
Ultimately the memory clock speed doesn't affect performance very much, so I'm not sure it matters.
It has sort of been that way in the past, so I wouldn't be surprised if there were similar issues with Ryzen at first.
However, I expect that there will eventually be some workstation grade motherboards from vendors like Supermicro which will be more direct about guaranteeing ECC support. This will be especially true if AMD decides to release some “official” small server grade CPUs on the AM4 socket (kind of like Intel’s Xeon E3 which uses the regular consumer socket, but requires a different chipset). I don’t know if there are any plans for such a CPU from AMD, but it wouldn’t surprise me considering there’s probably some real demand for something a little less expensive and powerful than the 32C/64T Naples server CPU. Rebranding a consumer 8C/16T Ryzen seems like a logical step, with the only cost being validation.
The thing to remember is that supporting ECC on an AM4 motherboard basically amounts to running a few extra traces between the CPU and DRAM slots, and adding support in the BIOS to initialize the memory controller properly. My experiencing in implementing boot code to initialize memory controllers leads me to believe that it’s probably one or two days worth of work to actually get it working (once everything else is done).
The real cost comes from testing and validation, which is why some manufacturers are so hesitant to give a straight answer regarding ECC support on consumer platforms. There is no point in spending the money to validate a feature that the majority of their customers think is a complete waste.
By offering “working but not validated” ECC support, they’re able to satisfy the demand of those who want to use these chips for small non critical servers without undercutting their higher margin server grade components.
ECC is a feel good feature, but is mostly useless. Memory either is corrupted (which can be detected by running MemTest) or it works perfectly fine without ECC. I have over 50 10+ year old Itanium servers with 192GB ECC RAM each, there are 2 ECC errors logged total, while running 24/7.
> Memory either is corrupted (which can be detected by running MemTest) or it works perfectly fine without ECC.
While there is reason to think soft errors are much less common than often claimed and therefore ECC memory has been much less necessary than people think (at least within the atmosphere), there's now a pretty good reason to prefer ECC memory because of rowhammer.
If you value the integrity of your data, particularly in longer term archive / recall / processing scenarios, then you value having ECC RAM as /part/ of a solution for detecting and correcting errors.
Not that it's likely to matter for these systems but there has been some recent work on using ECC to recover from errors caused by refreshing less often. This would use less power so I suspect we'll see someone try it in phones or laptops soon.
That's well below median of around 1+/year per server, and might indicate you are not logging all ECC errors found or just not using most of your memory heavily. However, ECC is much more useful for early detention of memory problems than actual recovery from memory problems. Further, if you get do large scale measurements you will get vastly higher numbers as the average is dominated by the worst offenders.
I run a few more modern servers and we see corrected ECC errors fairly often, but typically not randomly distributed across the boxes. This makes me suspect bad memory devices. We've replaced modules here and there with positive results.
Anyway as a former memory subsystem designer and a former employee of a memory chip maker, I like ECC. It provides some protection against bus noise errors even if you have full confidence in the memory chips themselves.
To word your statement differently: without ECC, your memory either works fine or is corrupted. But you don't know which one it is, until you run memtest. How often do you do that?
Personal anecdote: Upgraded RAM in my PC, ran memtest, everything fine. A year later, I need to move and so I consolidate all my data on a big hard disk. As the disk is new, I run some verification and notice some non-matching checksums. Long story short, the new RAM had become defective, and a new memtest showed a handful of errors after a few hours. Had I had ECC on that machine, I probably would have seen notifications about memory errors even before that day (plus the errors would likely have been corrected).
If Ryzen allows us to build a powerful PC with ECC for little more money than a non-ECC one, I'll happily pay for this "feel good feature". The current ECC-capable alternative from the Intel universe (Xeon CPU & motherboard) is simply too expensive for most people.
> To word your statement differently: without ECC, your memory either works fine or is corrupted. But you don't know which one it is, until you run memtest. How often do you do that?
Definitely this. The extra safety of mind offered by ECC RAM is totally worth the tiny price difference for me. The only thing that stopped me from actually using it was the lack of (affordable) CPU support. Ryzen changes that.
On that note, I hope it's not long until we can easily buy ultraportable laptops with ECC RAM without paying through the nose for Intel's ridiculous prices for their mobile Xeon CPUs.
What many don't realize is that ECC is a bit like paying for insurance just built in to hardware, most likely you won't need it, but if you can run ECC sticks other than cost there is no reason not to.
ECC is great for detecting errors more than correcting them. Then one can do thorough memtests and decide if it's time to replace memory modules. It doesn't really prevent problems (there are unrepairable ECC errors too) as much as it leads to early detection and repair. Well worth it when you're optimizing for MTTR.
It's not a silver bullet if the user is just paranoid about data integrity (nothing beats end-to-end integrity checks built into your system, while ECC only addresses one step in the process).
Personal anecdote. I have a Phenom + Asus mobo system with ECC memory in it that has been working fine for years without reporting a single ECC error. And then they started. At first there was a correctable error once every few days, then few errors per day, and then almost a constant stream of errors. Fortunately, they were all correctable. From the reported ECC syndomes I deduced that it was always the same single bit lane. That's why only a single ECC symbol was affected and the errors were correctable. I reseated the CPU and DIMMs and the errors went away, and the system keeps chugging on. Not sure what it was, maybe some dust particle getting in a wrong place, maybe some oxidation.
Anyway, I can only imagine how much trouble those errors could do if they were not detected and corrected. Especially, at the stage when there was a moderate number of them. I guess when they would become too many, then the system would just crash all the time.
Interesting! I had a similar setup using ECC memory on a GA-990FXA-UD3 with FX-8320. Unfortunately, what I discovered was, even though it operated OK, the actual ECC features weren't being utilized. What CPU / Motherboard combination did you use? I think it's exciting that your combination actually used the ECC functionality.
How much hardware have you used? I regularly find machine that crash, see errors reported in mcelog, and when I run memtest no errors. I'm guessing that memtest doesn't use a challenging access patterns, or something about powersaving/clocking that runs it a bit slower.
In any case I LOVE ECC. Sure random radiation caused bit flips are relatively rare. But it's REALLY nice to see high levels of ECC errors when something died and not have to wonder if it's the flash drive, power supply, graphics card etc. Sure sometimes its the dimm, sometimes it's the motherboard, sometimes it's the CPU. But with ECC errors you can much more quickly/easily tell if it changes when you swap things around.
>> If you're buying a desktop quadcore processor in 2017, you're doing it wrong. May as well buy a laptop.
Not true, IDEs, web browsers, office tools need single core performance and not 4+ cores. A laptop is more expensive, will make more noise when loaded, is harder to upgrade, is less customizable, has less ports, is more expensive to expand (ThunderBolt devices are not cheap at all). What are you talking about?
>> Intel's margins and lineup are utterly destroyed especially once you consider how expensive Intel HEDT boards are on top of it all.
There are few people who build workstations for themselves(even among software developers). HP and Dell haven't announced any Ryzen systems as far as I'm aware. Right now AMD's new chip will have minimal impact on Intel profits. Only if AMD can remain competitive on stability, performance and price for a few years will Intel start loosing major chunks of income and market share.
>> Kabylake sucks as a workstation chip but is ok for gaming if your PC is solely a glorified Xbox.
PCs are universal machines, you can play games AND do work on the the SAME PC ;-)
>> Vast majority (90%+) of gaming rigs are GPU-bound. Think very carefully before giving up 8 cores and 16 threads at the same price.
GPU-bound does not mean all CPUs produce the same framerates on a given GPU.
I personally do hate Intel and nVidia monopoly and hope AMD can increase their market share, but your points are very weak.
IDEs are very very easy to parallelise, since most of the hard work is done in background tasks, which can run simultaneously. Compilation also parallelises very nicely.
358 comments
[ 12.3 ms ] story [ 807 ms ] threadEDIT, in response to the other edit: sorry about that, I wasn't sure which comparison you were making. Still, the difference is slim enough that if I were in the market, I'd go with this one just to put the heat on Intel.
[1] http://tvtropes.org/pmwiki/pmwiki.php/Main/SeinfeldIsUnfunny (apologies for the TVTropes link at night!)
http://timmaurer.com/2012/01/16/horse-sense/
https://www.quora.com/Is-it-true-that-two-horses-pulling-tog...
I'm about to buy a beefy machine for FPGA development, I'm definitely considering AMD. Seems like I'd get better performance for a much lower price. I'm still waiting for more in-depth reviews though.
[1] http://core0.staticworld.net/images/article/2017/03/ryzen_ci...
Web and office? Single-threaded?
Why are these benchmarks, specially in 2018?
I do understand where you're coming from, but real-world performance is important, especially when that world is imperfect.
That's a load of bullshit. No one ever complained that the web is unbrowsable or an excel spreadsheet is unworkable without a high-end boutique processor. Those benchmarks are simply a load of bullshit that evaluate something that is utterly irrelevant.
With AMD's previous offerings, they benched well but ran like crap because so many applications are single-threaded. These single thread tests are important to know how these chips will work in real world scenarios.
This is the data enterprise buyers want to hear. If AMD can match Intel at a much lower cost, expect enterprise buyers to switch.
That said, those numbers aren't particularly bad for a new CPU running un-optimized software. All the important software will be updated soon enough once the compilers are in order.
https://www.pcper.com/reviews/Processors/AMD-Ryzen-7-1800X-R...
Perhaps you'd like to restrict yourself to 32 bit software just to be standards conformant?
If you do more than games, the Ryzen seems like a a good choice. It has serious power in applications, and is still fast in Games - which is a big step up compared to the FX. But no, it is not the goto choice for a gaming PC.
I'd take Intel CPU that runs 120 FPS average over AMD that runs 130 FPS average, when lowest FPS for Intel drops to 100 FPS and for AMD it drops to 80 FPS...
The Ryzen quads will come. I'm not at all sure what your point is here.
By ~15% compared to >160% for Bulldozer.
And this is day zero before any applications have been optimized for the new microarchitecture, compared with Core which everything has been targeting for more than a decade.
They did good.
Would you do it if you had the choice between that or adding new money making features/fixing bugs?
Of course I would, because people are going to benchmark my software against my competitor's software and choose the one that runs faster on their hardware.
Especially if my software is a compiler and people are benchmarking the programs it produces against other compilers, which is where most of the optimizing actually happens these days.
As for the priorities of the compiler devs, I think it will get some love.
Besides, if your application is one where performance is a serious issue, then you are going to spend some time optimizing rather than adding features. For some applications, performance is a feature.
Do you really care if your webpage renders for half a second longer?
If I was after encoding performance, I wouldn't buy a Ryzen. If I was after rendering performance, I wouldn't buy a Ryzen. Throw an mid-range nVidia on there for OpenCL, and the only CPU that makes sense for that price range is the 7700K.
If you're after pure performance, then you simply go with the best for your use case, whether it is Ryzen or not.
If you're on a tight budget, however, things get interesting.
hence, "things get interesting"
I expect if AMD went back through again, they could improve them even more.
(In the handbrake case, most likely it's half-speed AVX2 that is hurting them)
But why? http://www.anandtech.com/show/11170/the-amd-zen-and-ryzen-7-...
The benchmarks show Ryzen being the obvious choice other than cinebench single threaded.
Only if you are using a shitty software renderer or are completely stupid would you do rendering on the CPU.
The Ryzen performs nicely on the encoding tests, though that will be performed by the GPU soon. It already can be (NVENC and AMD VCE), but the quality isn't quite as good as everyone's favorite h.264 encoder, x264.
You're also missing that even on chrome the js and layout engines run different threads. Same with audio and video rendering. Same with downloads. The average page is probably running things on 10-20 random threads. And the average user has multiple tabs open.
Multicore CPU is so common these days that any performance critical code that can be parallelized, is. This includes GZIP and HTTP which are used heavily by web browsers. The network stacks underneath those in the OS are also multi core capable.
Single thread performance is not all that important anymore. It's better to have 2x the number of fast cores than half the number of slightly faster cores in almost every situation these days.
Not even. Video acceleration, browser acceleration, flash acceleration, and much more, are becoming increasingly GPU-oriented.
For the first time in a decade, AMD are seriously competing at the high end. That's great news for everyone except Intel.
Technically this is true, but unlike the desktop market where you can buy stuff in pieces, it really depends on how the laptop manufacturers adopt them. As in how fast they show up in product lines, and in which laptop segments they show up in.
If, for example, Intel gives laptop manufacturers pricing incentives (legally or not) to stick with their chips, it may not cause any problems at all with respect to actual customer purchases.
Cheap, power efficient, eight cores. It may not win everything, but it works for me.
It also has 1/2 cores, which doesn't show up on the single threaded benchmarks but explains the price and TDP differences.
The 7700k is losing in all aspects below w.r.t. the Ryzen 7 1800X, except absolute single-threaded perf. Intel is losing in:
- absolute multi-threaded performance
- relative multi-threaded performance per $
- memory bandwidth
- L3-cache bound workloads (AMD has twice more L3)
- price of motherboards
> - relative multi-threaded performance per $
Highly depends on the particular workload. Given the few benchmarks we have right now, it seems to go 50-50 once you factor in the lower price of the 7700k. At the very least, this one is close.
> - memory bandwidth
Seems about equal from the measurements I've seen (the theoretical numbers are irrelevant).
> - L3-cache bound workloads (AMD has twice more L3)
Cache-size bound workloads to be precise. Cache latency still seems higher, even on L3.
Overall, the 1800X seems to make the entire Broadwell-E series redundant on price/performance (at least at current price), no argument there. Against the 7700k, however, it's a very interesting decision, entirely based on the distribution of your workloads. And if Intel drops the prices on Kaby Lake, the value equation can change very quickly indeed.
http://www.guru3d.com/articles_pages/amd_ryzen_7_1800x_proce...
[1] https://www.pcper.com/reviews/Processors/AMD-Ryzen-7-1800X-R...
What's happening here is that AMD has priced Ryzen so inexpensively that it actually broke the usual market segmentation by driving high-thread-count workstation CPUs (usually $1k+) into the same price range as high-end gaming CPUs (~$300-400). Think of it as being able to buy an industrial dump truck for the same price as a pickup truck. The dump truck is clearly a better value in terms of hauling capacity, but most people would be far better off with a pickup truck for their daily traveling.
AMD's pickup truck (the 1400X) doesn't launch until later this year. The 7700K will probably still beat it on single-threaded workloads due to the difference in clock speed, but it should get you most of the way there for substantially less money, around $200 if the leaks are accurate.
Indeed. Single threaded performance, however, seems bad. Virtual cores most likely share less of the actual core than on Intel parts, allowing one to compete less with the other and get more done. On my i7 laptop, on anything more demanding, I see half the "cores" idling, probably because one is using the actual underlying core while the other is waiting.
Still, while AMD obviously did a good job, it will not make the Intel guys lose any sleep for now.
The server parts are another story. If they manage to reliably outperform Xeons (even if it's only on a per watt basis), it'll be fun to watch.
Right, thats exactly what happens with Hyper-Threading. If you can fully utilize the physical core, Hyper-Threading can't do much.
From the benchmarks, it seems a single thread can fully utilize all resources from an Intel processor while it can't on the new AMD ones and that's why their single-thread performance is lower and their multi-thread performance is higher.
We should keep in mind the benchmarks know nothing about the new AMD parts and how to keep all parts of the processor busy.
Yeah, as would be expected from most people on this site. Sad thing is, there's a bunch of pissed-off gamers right now who were on the hype train expecting Ryzen to be the next big thing in gaming, but it's far from it. Ryzen is great, just not amazing for gaming.
In so many years, I have never seen so much marketing where nothing actually exists in peoples hands, well since the introduction of 'segway', that took months of hype before we found out what it really was.
If it be true that this chip is what they say, then when will we see real benchmarks by real people? They're saying "March 5", yet the marketing stuff makes it appears that AMD already won, hell they're not even at the starting gate, hey I'm rooting for AMD, I wish this all to be true, but I'm not spending +$1,000 bucks for a new MB & Ryzen, until I know for a fact this thing is real.
Very interesting times in Silicon.
Something like that could be a monster for HPC & ML type tasks. Provided AMD gets their software and tooling up to par; Nvidia doesn't rule HPC/ML because their GPU's are leaps and bounds ahead of AMD, but because CUDA is..
If so, might explain why they went with a (relatively) gimped vector FPU for Zen compared to Intel offerings (particularly the soon arriving Skylake Xeons with 512-bit wide SIMD).
It seems that in common use (web browsing, office, gaming), fewer but stronger cores still shine and even the fastest Ryzen is slower and more expensive than Intel's offering.
So it really depends. A lot.
ECC should be a standard feature. If you don't want it you don't have to use it but to disable simply for market segmentation is lame.
Has Asrock explicitly stated that it works with all CPUs, or just that all their boards can do it (IF the CPU also supports it)?
> AMD Ryzen series CPUs support DDR4 2667/2400/2133 ECC & non-ECC, un-buffered memory
So the board seems to accept ECC memory, hopefully that means the memory controller actually performs ECC? I found a reddit comment claiming that the boards accept ECC memory but don't perform error correction[1]. Not a great source but seemingly possible.
Do you have any more sources suggesting that ECC is actually being implemented?
For example, BIOSs often have ECC settings, if we had some screenshots of an AM4 board bios showing ECC setting that would be strong evidence.
[0] http://www.asrock.com/mb/AMD/X370%20Killer%20SLIac/index.asp...
[1] https://www.reddit.com/r/Amd/comments/5v0cqo/ryzen_supports_...
In that case anandtech saying "We know that it is disabled for the consumer parts" and ASRock saying "ASRock AM4 motherboards fully support ECC function" can both be true.
https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
This is set up by the firmware (BIOS). Since we all already know that quite some of these still have issues (unsurprisingly, it's a new platform), it isn't far-fetched at all that current firmware doesn't quite do the right thing with certain memory sticks.
Note that this is unbuffered, unregistered ECC memory, which is usually more expensive and less available than the standard server memory (DDR3R / DDR4R), which is NOT compatible with any consumer CPUs.
Right now the #3 system in the world is a Cray system using Opterons.
Edit: If no motherboard supports its, in Linux one can use ecc_enable_override (similar thing might be available in Windows). That is if the CPU support it. It can USE ECC memory, but we don't know if it can use the ECC checksums, most motherboards have a NO here.
Might as well ask her about the possibility of coreboot if we are at it.
Edit: Account made, instructions received. If all goes well I should be asking later today.
https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
btw: someone asked about coreboot: https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
http://webcache.googleusercontent.com/search?q=cache:wO1MKok...
I agree ECC should be a standard feature, but it follows the same marketing patterns (in that industry) as disabling multi-processor support or artificially limiting clock speeds.
Without ECC you may get a completely wrong answer when the increasingly small, vulnerable, and numerous dram cell is flipped.
I don't like this situation, but I certainly understand it. AMD needs the server market to raise revenues. This means its server offerings, from soup to nuts, need to be rock solid. Letting whitebox resellers junk up their reputation is not within their interests.
Not to mention, enterprise sales often subsidize consumer sales. Those big enterprise markups go into the same pot of money that allows AMD to sell consumer chips at lower prices. Anything that threatens enterprise sales could affect the consumer space. AMD needs to be very careful here as, compared to intel, it needs to prove the value of its brand every chip generation. They also need to overcome a lot of bad mojo they've developed from questionable benchmarking and marketing of its previous two, or more, generations of chips. Bulldozer, I believe, had half the single core performance compared to the intel equivalant at the time, yet it benchmarked well due to its extra cores. The market doesn't forget shenanigans like these.
That said, the market allows 'workstation' level cpus and motherboards that support ECC. A little legwork can find you decent deals, but no, they're not going to be $59 NewEgg specials and you won't be able to use the consumer Ryzen on it. Consumer CPUs are almost always gimped anyway - smaller caches, disabled features, slower clocks, etc. The ECC limitation isn't too different from this, but it is annoying.
Hopefully more software starts to make use of multiple cores more effectively. It's not like it's a new feature of chips any more.
http://www.intel.com/content/dam/www/public/us/en/documents/...
tomshardware: http://www.tomshardware.com/reviews/amd-ryzen-7-1800x-cpu,49...
Also, another AMD employee chimed in:
"But what's also clear is that there's a distribution of games that run well, and a distribution of games that run poorly. Call it a "bell curve" if you will. It's unfortunate that the outliers are some notable titles, but many of these game devs (e.g. Oxide, Sega, Bethesda) have already said there's significant improvement that can be gleaned. We have proven the Zen performance and IPC. Many reviewers today proved that, at 1080p in games. There is no architectural reason why the remaining titles should be performing as they are." Source: https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
Unfortunately you can bet Intel is going to put all the spin they can into "AMD underperforms for gaming", while the software game developers are going to slowly optimize for Zen, up to the point where this perf discrepancy will be fully fixed but people will still have the mindset that "AMD unperforms for gaming"... sigh
http://www.agner.org/optimize/blog/read.php?i=49#49
https://software.intel.com/en-us/gamedev/
But also things like physics engines (eg. Havok, which was acquired by Intel and heavily optimized for Intel processors), etc.
And of course Intel's C Compiler is known to cripple AMD: http://www.agner.org/optimize/blog/read.php?i=49 And some/many games are compiled with ICC...
Honestly I'm not very optimistic about that happening. Maybe some will. But the mentality these days seems to be that there's one winner and fuck all the rest -- we're busy and we don't have time to optimize for losers. Like with web developers seeming to almost exclusively work with Chrome except where they're told to support IE (or Safari?). Firefox just isn't popular.
Still, there's a chance the multi-core revolution will eventually happen. Vulkan could help us get there, although weren't DX12 was supposed to provide similar improvements in multi-threaded access to the GPU?
The fix is to have a driver to help Windows treat a CCX almost as if it was its own socket. See https://forums.anandtech.com/threads/official-amd-ryzen-benc...
In theory a workaround would also be to disable cores 4-7 (and keep 0-3) in the BIOS. Sadly no Ryzen reviewers tried it...
Compiler tests sound more interesting to me, I'm looking forward to those benchmarks.
Boards initialization takes forever
For the last weeks Computerbase had access to different mainboards [...]. They had an extremely long init process in common: even with the latest BIOS it took up to 30s to get to the POST screen. Obviously this needs optimization.
-----
German:
Die Boards initialisieren ewig
ComputerBase konnte die letzte Woche auf drei verschiedene Mainboards zugreifen, neben den X370-Varianten von Asus und MSI auch auf das Gigabyte GA-AB350-Gaming 3 mit Chipsatz B350. Allen gemein war auch mit dem letzten BIOS der extrem langwierige Initialisierungsprozess: Bis der Post-Screen erscheint, vergeht über eine halbe Minute. Auch hier gibt es offensichtlich noch deutlichen Optimierungsbedarf.
----
https://www.computerbase.de/2017-03/amd-ryzen-1800x-1700x-17...
But the actual compilers people use (IE not that version of GCC) are rarely well-optimized for CPU. In fact, most compiler profiles are not flat yet, and where your compilation takes a long time, it's because some optimization or algorithm has gone nuts and needs to be fixed.
Until that kind of state of the world changes, measuring compile time is IMHO, a bit pointless, because they are only artificially CPU bound, and not in a way that is useful to benchmark (IE while i understand that people use these compilers every day, and care about performance, i'm just saying that, given a compiler benchmark, i could still probably make it take zero time with limited amounts of work)
Sure, in that sense, but you also don't want benchmarks that are trivial to game by performing optimizations to them. :)
If someone managed to make an optimization that could suddenly make GCC 2x faster without humans editing GCC's source, that would be a major CS breakthrough :)
However,
"If someone managed to make an optimization that could suddenly make GCC 2x faster without humans editing GCC's source, that would be a major CS breakthrough :) "
No, it really wouldn't. For example, if you optimize the line-ending and tokenization finding in GCC's parser to use SIMD, you would become significantly faster. Not 2x, but significantly. These are entirely doable by compilers. " But you don't ever get to re-write the benchmark itself, you can't replace a 2^N path with an N^3 path or anything like that."
Except, you can, by hoisting calls out of loops, etc. There are also dynamic memoization optimziations that can be applied, etc.
"If someone managed to make an optimization " This is a common fallacy that it's ever really one optimization. It's usually combinations of optimizations, each buying you N%.
I will state outright, having worked on pretty much all of the slow parts of gcc (according to spec profiles), and having a team of 80+ people working on compilers, 2x on gcc through compiler optimization is more than within the realm of possibility.
For some reason one of my sideprojects needs more than two minutes to link but the task manager tells me it's neither CPU nor RAM nor disk bound. I haven't found the cause yet.
http://imgur.com/a/vDHAv
It looks like Intel's retail channel partners have a lot of inventory still in the pipeline. Prices are still within the bounds of the second half 2016. My best guess is the retail channel partners will be the most hurt by any slowdown of Intel CPU sales.
[1] http://www.digitaltrends.com/computing/intel-cpu-prices-drop...
http://www.microcenter.com/product/472529/Core_i7-7700K_Kaby...
http://imgur.com/a/vDHAv
AMD is trying to change that. Of course 16t games won't be available from day one of Ryzen launch.
Ryzen is much better positioned anyway because it has almost comparable IPC (Broadwell levels), that means even when you get a dual core/single core optimised game you'll get good performance rather than the horrible situation with older AMD CPUs.
http://store.steampowered.com/hwsurvey
What some people here are anticipating is the situation where games really do scale well to 8 cores, at which point we'd likely see the Ryzen compete against Intel's stupidly expensive 8-core parts and roughly match them in performance while solidly beating them in price/performance ratio. Just like in the encoding benchmarks that actually do run well with multiple cores. These games aren't here now. We aren't there now. But it might be the future.
Would be nice to make a simple one-click benchmark: download a VirtualBox VM and click Start — it boots mfsbsd and starts compiling to a RAM disk, measuring the time.
I'm going to build an 1800X-based box in a few weeks (for games; planning to run Steam on a GNU/Linux distro), I'll want to see `make -j16 buildworld` for sure.
http://www.tomshardware.com/reviews/amd-ryzen-7-1800x-cpu,49...
https://translate.google.com/translate?sl=de&tl=en&js=y&prev...
The 1800X does seem like a killer value for the money for us C++ people :)
You can get Radeon R7 250 under $50, which absolutely destroys i7-6700k's integrated HD530.
(Technically HDMI 1.4 also supports 4K but only at 24hz, which is fine for TVs but a miserable experience on a PC desktop)
Hence my question about having GPU integrated to the motherboard like older boards had it.
If neither monitor has HDMI 2.0 then an RX470 would be required to get dual DisplayPort outputs.
3.6 GHz (4.0 GHz Turbo) over 16 threads
15,812 Passmark for ~$500
-----
Intel i7-7700K (Kaby Lake):
4.2 GHz (4.5 GHz Turbo) over 8 threads
12,321 Passmark for ~$350
-----
Intel Core i7-6850K (Broadwell E):
3.6 GHz (4.0 GHz Turbo) over 12 threads
14,500 Passmark for ~$575
-----
Competition is back. Ya!
It even has a name '"buy the rumor, sell the news"
http://www.investopedia.com/terms/n/news-trader.asp
The idea behind this, simply put, is that if you are in a trade for whatever reason with a stock that has a major announcement coming; you are pretty much at the potential peak at announcement time so it is a good time to get out.
https://www.purepc.pl/procesory/premiera_i_test_procesora_am...
Maybe knowing consumers will probably have more than 2 cores will make games optimise better for parellelism, it's not like the engine programmers _cant_ do it, they do it all the time with GPUs.
When making AI bots it's quite common to have all of the work of processing the AI it's own thread, so each rendered frame will do all calculations for all bots. That in of itself would be cheap, except it has to share it's core with a bunch of other crap.
With 2 cores you can have a main thread for updating the screen and then an "everything else" thread. If you can't be sure people will be playing with >2 cores then you have to keep your budgets down to 2 cores when developing and that's quite limiting.
it's not that it's "hard" to parralellise(?), it's that you run out of resources super quick if you make assumptions about CPU topology.
https://www.purepc.pl/procesory/premiera_i_test_procesora_am...
A Broadwell with 3300 Mhz (3700 Mhz boost) beats a Kaby Lake (Skylake...) with 4200 Mhz (4500 Mhz boost)? Can the Broadwell use the 128MB EDRRAM from the iGPU as a L4 cache if the iGPU is not used? Would that actually make such a difference?
That's exactly what it does, and it gives you a "free"* ~20% increase in memory operations performance (which tends to be a ~20% increase in general performance).
* not actually free because you paid for those 128MB EDRRAM...
But from a consumer point of view, it increases performance (either for graphics or for the whole system if I have a discrete GPU), and it can also help with GPU numerical processing, because you get "free" transfers from memory (L4) to graphics memory by flipping a bit (or a few bits) in hardware. Don't know why this wasn't expanded further, appart from the cost of the embedded RAM.
http://www.anandtech.com/show/10968/the-intel-core-i7-7700k-...
https://www.techpowerup.com/226719/amds-zen-to-implement-adv...
It's possible only Naples will support it.
Interesting would be 4C/8T parts - if they can run them cool (given 8C/16T 1700 uses 65W I think it's not unreasonable), they probably can clock them to get to 7700k parity at 95W.
It's a pity that they did not include at least 32 lanes on-chip to allow two full x16 GPUs at least on the 1800X, but I hope there is a 1900X in the pipeline with more lanes.
[1] https://arstechnica.com/gadgets/2017/03/amd-ryzen-review/
http://www.anandtech.com/show/11170/the-amd-zen-and-ryzen-7-...
7th) Will all Zen products have all of the instruction sets and platform extensions, or could lower end chips lose features like virtualization?
A7: In the consumer client space we have no plans to turn off virtualization or features.
9th) Does AM4 / consumer ZEN support ECC memory?
A9: ECC is enabled on Ryzen and AM4.
[https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...]
edit: added link
Don't get me wrong, I have a bunch of rack pull servers and they're great for the price but I colo mine for less/server than the power bill/server would be at home and have them ona faster connection. I stick with low power options for light in-home usage
Modern processors do some clever tricks to minimize using power if it is not needed, and have gotten progressively better at it. For home use cases, it's often worthwile to spend more money to get newer or otherwise more efficient PC components. This is especially true in areas where electricity costs are higher (e.g. average price per KWh can be almost twice as high in the Benelux when compared to the US).
One thing to get an indication about how good a CPU/chipset is is idle power usage of a system. It seems Ryzen is even better at this than Intel's already excellent Sky/Kaby Lake architectures [1], possibly because of the fact that Ryzen motherboards contain fewer components. On the contrary, the Ivy Bridge architecture of the chip mentioned by the parent cannot be considered competitive from a power consumption perspective, despite still being a great performer.
[1]: https://www.pcper.com/reviews/Processors/AMD-Ryzen-7-1800X-R...
https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
[0]https://youtu.be/9wJQEHNYE7M?t=4m43s
Ultimately the memory clock speed doesn't affect performance very much, so I'm not sure it matters.
From what I've been reading about the upcoming Ryzen, no motherboard so far claims to support ECC.
However, I expect that there will eventually be some workstation grade motherboards from vendors like Supermicro which will be more direct about guaranteeing ECC support. This will be especially true if AMD decides to release some “official” small server grade CPUs on the AM4 socket (kind of like Intel’s Xeon E3 which uses the regular consumer socket, but requires a different chipset). I don’t know if there are any plans for such a CPU from AMD, but it wouldn’t surprise me considering there’s probably some real demand for something a little less expensive and powerful than the 32C/64T Naples server CPU. Rebranding a consumer 8C/16T Ryzen seems like a logical step, with the only cost being validation.
The thing to remember is that supporting ECC on an AM4 motherboard basically amounts to running a few extra traces between the CPU and DRAM slots, and adding support in the BIOS to initialize the memory controller properly. My experiencing in implementing boot code to initialize memory controllers leads me to believe that it’s probably one or two days worth of work to actually get it working (once everything else is done).
The real cost comes from testing and validation, which is why some manufacturers are so hesitant to give a straight answer regarding ECC support on consumer platforms. There is no point in spending the money to validate a feature that the majority of their customers think is a complete waste.
By offering “working but not validated” ECC support, they’re able to satisfy the demand of those who want to use these chips for small non critical servers without undercutting their higher margin server grade components.
Link on reddit: https://www.reddit.com/r/Amd/comments/5x4hxu/we_are_amd_crea...
While there is reason to think soft errors are much less common than often claimed and therefore ECC memory has been much less necessary than people think (at least within the atmosphere), there's now a pretty good reason to prefer ECC memory because of rowhammer.
http://thememoryguy.com/using-ecc-to-reduce-power/
PS: Radiation also makes a huge difference.
Anyway as a former memory subsystem designer and a former employee of a memory chip maker, I like ECC. It provides some protection against bus noise errors even if you have full confidence in the memory chips themselves.
Personal anecdote: Upgraded RAM in my PC, ran memtest, everything fine. A year later, I need to move and so I consolidate all my data on a big hard disk. As the disk is new, I run some verification and notice some non-matching checksums. Long story short, the new RAM had become defective, and a new memtest showed a handful of errors after a few hours. Had I had ECC on that machine, I probably would have seen notifications about memory errors even before that day (plus the errors would likely have been corrected).
If Ryzen allows us to build a powerful PC with ECC for little more money than a non-ECC one, I'll happily pay for this "feel good feature". The current ECC-capable alternative from the Intel universe (Xeon CPU & motherboard) is simply too expensive for most people.
Definitely this. The extra safety of mind offered by ECC RAM is totally worth the tiny price difference for me. The only thing that stopped me from actually using it was the lack of (affordable) CPU support. Ryzen changes that.
On that note, I hope it's not long until we can easily buy ultraportable laptops with ECC RAM without paying through the nose for Intel's ridiculous prices for their mobile Xeon CPUs.
It's not a silver bullet if the user is just paranoid about data integrity (nothing beats end-to-end integrity checks built into your system, while ECC only addresses one step in the process).
In any case I LOVE ECC. Sure random radiation caused bit flips are relatively rare. But it's REALLY nice to see high levels of ECC errors when something died and not have to wonder if it's the flash drive, power supply, graphics card etc. Sure sometimes its the dimm, sometimes it's the motherboard, sometimes it's the CPU. But with ECC errors you can much more quickly/easily tell if it changes when you swap things around.
- If you're buying a desktop quadcore processor in 2017, you're doing it wrong. May as well buy a laptop.
- Kabylake sucks as a workstation chip but is ok for gaming if your PC is solely a glorified Xbox.
- Vast majority (90%+) of gaming rigs are GPU-bound. Think very carefully before giving up 8 cores and 16 threads at the same price.
- Intel's margins and lineup are utterly destroyed especially once you consider how expensive Intel HEDT boards are on top of it all.
- AMD made CPUs Great Again. Bravo AMD!
[0]https://videocardz.com/66826/amd-ryzen-7-review-roundup
[1]https://www.youtube.com/watch?v=9wJQEHNYE7M
Not true, IDEs, web browsers, office tools need single core performance and not 4+ cores. A laptop is more expensive, will make more noise when loaded, is harder to upgrade, is less customizable, has less ports, is more expensive to expand (ThunderBolt devices are not cheap at all). What are you talking about?
>> Intel's margins and lineup are utterly destroyed especially once you consider how expensive Intel HEDT boards are on top of it all.
There are few people who build workstations for themselves(even among software developers). HP and Dell haven't announced any Ryzen systems as far as I'm aware. Right now AMD's new chip will have minimal impact on Intel profits. Only if AMD can remain competitive on stability, performance and price for a few years will Intel start loosing major chunks of income and market share.
>> Kabylake sucks as a workstation chip but is ok for gaming if your PC is solely a glorified Xbox.
PCs are universal machines, you can play games AND do work on the the SAME PC ;-)
>> Vast majority (90%+) of gaming rigs are GPU-bound. Think very carefully before giving up 8 cores and 16 threads at the same price.
GPU-bound does not mean all CPUs produce the same framerates on a given GPU.
I personally do hate Intel and nVidia monopoly and hope AMD can increase their market share, but your points are very weak.