One of the biggest improvements to the Ryzen 7 X3D series (e.g. 9800) is that the caches have been moved from one side of the die (to the other), which places the major heat source closer to the heat sinks.
Power is heat, and the amount dissipated is the square of the voltage- a processor that is reliable at less voltage means you can get a lot more frequency in the same heat envelope.
Yet reliability decreases as frequency goes up, and that can only be stabilized by adding more voltage- so the faster you run the processor, the more voltage you ultimately have to give it, so the power/heat produced grows exponentially until you can't get rid of the heat fast enough (at which point your only option is to actively cool the chip).
This is overclocking 101.
Note that classic overclocking was viable because of arbitrage- buying a processor, pushing it to the point it got too hot, and stress-testing it at that temperature to ensure reliable operation. Processor manufacturers all do that formal verification at the factory now as they'd be uncompetitive otherwise, especially in laptops.
2 years sounds very fast. It seems to be because the 3d cache SKUs of each generation were released at different stages. The Zen 3 one was a later variant.
First zen 3 November 5, 2020 with desktop processors. First desktop Ryzen 9000 processors on August 8, 2024. So the generations were about 4 years apart.
Yes I stated that as well in the previous comment [1] on how Apple got 50% faster in 3 years.
>Compared this to the article "How did AMD Ryzen get 50% faster in two years?" The Zen 3 uArch being used in the article came out in 2020. So it is more like AMD got 50% faster in 5 years.
I was wondering where you got the info about A20 Pro: "shrink back to 9 decode width". Also I imagine the A20 Pro is very similar to an M6, but downsized.
Not implying anything here, but my comment was actually before yours, this article was reposted via HNs 2nd chance pool, https://news.ycombinator.com/pool
I suspect the improvements are even more dramatic going from Zen 1 through to Zen 5. AMD has really hit the jackpot with how scalable the Ryzen CPU is considering how they're able to improve the performance from year to year. This is a stark difference to the FX series during the 2010s, which saw very small YoY performance increases by comparison. Ryzen really is AMD's equivalent to what Nehalem/Core was for Intel back in the mid 2000s.
That is all true but I will defend the FX series a little. Mostly now that there is a lot of software that scales across cores better now, they haven't aged as terribly as others have. They aren't great but not terrible considering.
>Mostly now that there is a lot of software that scales across cores better now
That's pretty much irrelevant since the AMD's FX (Bulldozer and Piledriver) arch's issues weren't that SW at the time wasn't using all the 8 cores. Intel dropped the Core 2 Duo and Quad into the era where most SW was still single threaded and those CPUs still ripped regardless.
Here's the big reasons why the FX sucked back then and why they still sucks today in the multicore SW era:
Instead of discrete, fully independent cores, AMD grouped processing units into "Modules" where each module contained two integer execution units, but they had to share critical resources like one FPU, the instruction fetch/decode pipeline, and the L2 cache so when both "cores" inside a module were heavily taxed especially with math or physics-heavy calculations, they choked fighting over shared hardware.
AMD designed Bulldozer with a very long pipeline, betting they could sacrifice efficiency per clock cycle in exchange for extraordinarily high clock speeds but the IPC was so bad that an FX core was often slower clock-for-clock than AMD’s previous-generation Phenom II chips and also power consumption exploded.
FX processors were plagued by high cache latencies and an inefficient memory subsystem
So these are the main bottlenecks today regardless of parallelization of desktop SW.
I remember rumblings a decade-ish ago that basically their hand was possibly forced to release the thing to avoid a full on revolt about abandoning all of their work; after all, Intel had been investing in deep pipelines for a while before, they were strapped for cash after the ATI Acquisition, and other 'server-ish' CPUs had done CMT type things in the past (keeping in mind that AMD was seeing a huge surge in server market share due to Hammer.)
Hyperthreading/SMT is a significant boon for heavily threaded workloads. What makes that such a win while Bulldozer's implementation of "two integer units sharing a front-end, cache and FPU" is supposedly so bad? Because that description makes the 8 core Bulldozers sound exactly like a 4 core with SMT.
AFAIR Steamroller was a big 'correction' of the Shared resource issues in the arch (I can't remember if other revisions had other improvements).
AMD was also having to deal with the fact GloFo split off and was relying more on general 'bulk' lithography, which kneecapped them for some time especially due to yield issues on the FX series and overall cost of that deal.
Intel also very quickly after, released Sandy Bridge and aggressively scaled it up and down; the 2500K was so cheap yet powerful I know of at least one setup that ran for a decade an only got replaced because they needed to upgrade to windows 11 for compliance-esque reasons. My own 2500K I replaced in 2017-2018-ish, only because either the motherboard took an unfortunate dive and it was easier to replace both at once.
FWIW, I did do a cheapie FX build in 2015ish for my then-girlfriend as a DVR and light gaming/emulation 'under the TV box', and it did the job well for the price, but it definitely wasn't anything amazing.
It was a tough time for AMD for sure. I think the 'split' between the Cat cores (Bobcat/Jaguar) also hurt them from a resource standpoint, although one could argue that it also kept them alive to recover (i.e. Jaguar in XBox One and PS4 being a volume contract part) [0]. They did a lot of moves that caused short term pain (that glofo spinnoff helped pay off the ATI Acquisition AFAIR) but helped them become the company that is still surviving today.
[0] - One odd side note, I still find it odd that they never did a dual channel Jaguar laptop part. I still ask whether it was because it would have made the FX look that bad...
> Instead of discrete, fully independent cores, AMD grouped processing units into "Modules" where each...
Yeah that's hyper-threading intel was doing it as well and all modern CPUs do it as well. Where AMD dropped the ball, was they did not disclose that in their marketing as clearly as they should.
All CPUs today are marketed as x cores 2x threads, back then some AMD marketing genius in their infinite wisdom put 8 cores on the box, instead of the honest 4 cores with hyperthreading.
No. it's just the opposite of Hyper-thereading.
SMT it's about maximize the resource usage of a CPU, running 2 or more threads at the same time.
AMD used the opposite technique of SMT. It had a technical name that I can't remember now, and wasn't invented by AMD.
>Instead of discrete, fully independent cores, AMD grouped processing units into "Modules" where each module contained two integer execution units, but they had to share critical resources like one FPU, the instruction fetch/decode pipeline, and the L2 cache so when both "cores" inside a module were heavily taxed especially with math or physics-heavy calculations (like in videogames), they choked fighting over shared hardware.
They did just fine in parallel workloads, so I think this is not accurate. The design scaled just fine. The problem was that each core was weak.
>They did just fine in parallel workloads, so I think this is not accurate
They in fact DID NOT "do just fine in parallel workloads". The contemporary competition from Intel that was 4-core + SMT was beating AMD's 8-core FX CPUs in plenty of multi core tasks. AMD only won in >4-thread strictly integer benchmarks.
When we say a design "scales", that means that increasing the size of the workload does not incur a lot of overhead. If contention between shared resources meant that the design was not able to achieve an ~8x speedup when run with eight parallel threads, that would mean the design was not scalable. But we did in fact see a roughly 8x speedup with eight threads, so the design scaled just fine. The problem with the design was that each core was individually crummy, so even eight cores running in parallel had lackluster performance.
The myth that each two-core module functioned more like one core with hyperthreading would suggest that these CPUs would have much higher per-core performance when lightly loaded than when fully loaded. That is not what happened. Each core was crummy even when lightly loaded, but under full load you would have eight crummy cores, which would beat four Intel cores on a lot of workloads.
The only time contention was a serious problem was with workloads that were dominated by floating point, which were relatively rare.
> If contention between shared resources meant that the design was not able to achieve an ~8x speedup when run with eight parallel threads, that would mean the design was not scalable.
By that definition it definitely was not scalable.
>But we did in fact see a roughly 8x speedup with eight threads,
Care to share a source? Because AFAIR there definitely was no 8x linear speedup with 8 threads even in benchmarks, let alone in real world use cases. The only benchmarks where those 8 threads would scale best and beat Intel were archival compression/decompression and media encoding. At everything else Intel wiped the floor with it.
>The only time contention was a serious problem was with workloads that were dominated by floating point, which were relatively rare.
Many real-world compute workloads, especially gaming related, are floating point.
>Instead of discrete, fully independent cores, AMD grouped processing units into "Modules"...
This argument has never made very much sense to me. Yes, the decision to group cores into modules with some shared resources did introduce a bottleneck and result in lower performance than having isolated cores. But this didn't make the processor worse than if it didn't have the additional cores at all. These processors were at their best on highly parallel workloads, despite the presence of this bottleneck. AMD shipped far more cores than Intel at the same price and as a consequence came out ahead on highly parallel workloads. The contention for the shared front end was not serious enough to make up for the core count advantage. How the module architecture an explanation for why the design failed, when the bottleneck only becomes relevant in situations where the design is winning?
It's hard to find benchmarks from 15 years ago. But Phoronix finds what I describe: the design doesn't scale quite as well as a "true" eight-core design, but still scales better than the competition of the time which had lower core counts. Though it does seem straightforwardly bad at parallel floating point workloads.
It's as FlowingRiver said. If we were to release the same parts again with modern software, Bulldozer would be a far stronger competitor to Sandy Bridge. FX aged better than Intel's designs of the same era. But Intel's designs were better to start with, so I'd still take the Core.
The reason they sucked is that they ran hot and their per-core performance was terrible.
I did run a multithreaded CPU only n-body problem resolvers in FX cores. So heavy double precision work and putting each core to 100% .
A Fx-8370E would get a speed up around 7.5 times Vs the single threaded version
A FX-4300 would get a speed up around 3.9 times Vs the single threaded version
So in a physical heavy computation task, involving double precision math (where the AMD design of share FPU units should penalize most), the FX cores where happy churning numbers with the expected speedup Vs a single core/thread version of the code. So stop saying that FX cores sucks at multithread. They fucking worked fine on that kind of tasks.
I just was to say the same thing. I had a good experience with a FX-8370E.
They go for too many cores to early, and sacrificed some of the CPU performance to do it. The gamble gone wrong...
Yes, my FX 8350 was doing really well on my Gentoo machine for the price. It really depends on the use case - AMD went hard in on multi core while Intel focused on single thread perf.
Part of the reason we didn't see much in the way of YoY improvements for Bulldozer, is that AMD almost immediately abandoned it and threw resources at Zen after it launched.
Steamroller did see 30% IPC improvements over Bulldozer (all the design work would have been done before they switched to Zen), but AMD canceled the full FX version, and only ever shipped the APU version of Steamroller (with only 2 modules, aka 4 threads).
If they had shipped a Steamroller FX cpu, the generational improvements would have looked similar to many of the generational improvements that Zen received... but didn't really matter as Bulldozer started so far behind.
They never shipped the zen 5 epyc big cache chip (e.g. 9685X). :( I'd hoped they pulled it to produce a HBM integrated chip instead, but they didn't do that either.
I don't think there are many fabs capable of producing HBM
My hope is on Huawei, which has recently developed a stacked chip architecture that reminds me of HBM, and I'm wondering if that's step 1 of China entering HBM production.
I don't know enough about hardware, so I might be saying absolute nonsense, happy to be corrected so I can learn where I'm wrong
I looked at game benchmarks, and sometimes 5600 or 5600X were fastest. With all X3D processors so fast that I would not decide to upgrade above 5700X3D or anything cheaper. Now, with current RAM pricing, getting a X3D CPU with 90+ MB L3 Cache seams the way to go instead of doing a DDR4 -> DDR5 leap.
Maybe I should do the same. I have 3800XT, I could go to 5900XT for like $300. I don't see myself upgrading anytime soon because these prices are laughable and will eventually come back down when the AI bubble pops.
I feel like most benchmarks don’t really capture CPU benefits to gaming.
It’s even difficult to find CPU benchmarks that don’t overemphasize 1080p and eSports scenarios.
My upgrade from 5600x3D to 9850x3D felt kind of dumb at the time, but I decided to do it because Micro Center’s bundle deals are so far below market pricing.
I was shocked at how much better it is. Benchmarks and FPS don’t really show things like micro stutters and little performance wrinkles like that. I’m not even sure 1% low FPS counts capture it.
A great example game for this is Oblivion Remastered. Upgrading my CPU alone with the same GPU took away the environment loading slowdown almost entirely.
The benchmark will tell you that I didn’t gain any FPS during gameplay but every time I open a door into the new environment my CPU is positively impacting the experience.
I would have said the exact same thing you are saying until I actually experienced upgrading to the best on the market. For the record, this is the first time in my life I’ve actually owned the best CPU on the market for gaming.
My old advice would have been to buy one of those sweet spot cheaper mid-range gaming CPUs, but my newer advice is really if you’ve already spent all that you’re willing to spend on a GPU (I have a 9070XT, my only upgrade paths are insanely expensive), buy the highest gaming CPU on the list for gaming benchmarks (e.g., I wouldn’t go crazy with a 9950X3D2 since it doesn’t have any gaming improvements above the 9850X3D).
And the thing about CPUs is they’re not insanely expensive like GPUs. We are talking a price delta of $200 between this beast of a CPU and something way more middling.
Oblivion remastered is an absolute dumpster fire for performance in general.
Given the massive increase in hardware prices lately, we are going to have to get by with mid range or even low end hardware for a lot longer. Studios will be forced to actually optimise games to not run like shit on a sub $8000 PC.
Every game should be targeting the switch 2 and steam deck in terms of power.
> Every game should be targeting the switch 2 and steam deck in terms of power.
I think PS5 is a perfectly reasonable spec to be targeting, which was the higher end of mid range about 5 years ago, and still significantly out performs Switch 2 (which significantly out performs steam deck). I’m sure the GTA sales will demonstrate that the install base at this spec is wide enough for broad commercial success.
I don’t really want game experiences to be locked at a PS5 forever. That console is actually somewhat old now. There are phones that have more graphical power than a PS5. A MacBook Air has a more powerful GPU than a PS5.
I grew up with gaming systems that got rapidly better and while it can be argued that we are at a diminishing returns plateau of gaming horsepower, I don’t think the solution is to never get hardware upgrades ever again, even in the current RAM-constrained environment.
I didn’t mean to imply that we should target ps5 as an exclusive platform, but ps5 is a good hardware spec to target, as there are several platforms that can match it and it has a wide install base.
Yeah, and I see what you mean there. I think even when the PS6 comes out, game developers largely won’t be taking advantage of its capabilities just like how today they don’t really leverage the capabilities of an RTX 5090.
We saw with much of the PS5’s lifespan that many of its cross-platform games came out on PS4 and original Switch with minimal compromise.
Not enough of the install base owns those higher end solutions, so for the most part someone with a better performing rig is getting better FPS, higher resolution, and not much else. Even gaining additional draw distance is rare these days, and higher resolution textures are basically impossible to see.
If the console is going to survive rather than decline as a whole I think that Microsoft and Sony will need to get more serious about innovation rather than just making a samey spec bump at a high price that doesn’t motivate their users to upgrade. Nintendo already proved that basic and rather mild level of innovation can excite the market (original Switch).
Actual gameplay capabilities haven't been limited by hardware for a long time now. You can implement any mechanics you want on modern hardware, the only thing limited is graphics, resolution and framerate.
It's relatively easy to allow powerful hardware to just run the game at 4k 120fps highest settings while weaker hardware runs exactly the same game at lower graphic settings.
But that’s the thing: I want to play oblivion remastered.
Having a powerful rig is the brute force way around the game’s performance problems.
My PC wasn’t $8000. My 9850X3D, 32GB of RAM, and motherboard cost $800 in a bundle. Yes, purchased in 2026 during the RAM crisis. I was also able to sell my previous CPU, RAM, and board for around $500.
Obviously this is not a budget build but about half the cost is the graphics card. I would have saved minimal amount of money if I had gotten a worse CPU.
Most benchmarks use singleplayer games with built it benchmark, you'll not see the biggest difference there. If you play online games like shooters, MMO, anything really with multiple players or entities, theres usually a much bigger difference. Also in rts games and such with many units.
> It has the highest base clock of any modern AMD chip (4.3 GHz), plus 16-cores, which is already more than most workloads need.
Does the base clock even mean anything beyond something that may or may not vaguely indicate the performance? Its lower than what these chips can sustain with sufficient cooling from my experience. It's higher than what the chips will clock down to when not sufficiently cooled.
Base clocks from AMD are nothing more than the guarantee for any workload when the system exactly meets the advertised power/temperature requirements.
They aren't the maximum or minimum but I think they are the only clocks with a specific guarantee "if you match X, you must be able to get Y". The max boost clocks usually aren't too hard to reach if you're trying though, there is just no guarantee you will in all workloads or how much it takes to get there.
I have the laptop version 9955HX3D (which I believe is a 9950 with 2.5ghz base, 5.5 boost, and lower TDP) in a Legion Pro 7 and it's a fantastic chip for a computer that I can move around. 16 full cores in a laptop is a luxury and 64gb of ram has proven to be a wise choice prior to price spikes
Let's not discuss battery life and overall size, however it helps the thermals.
I do a lot of CPU bound engineering and scientific compute on the computer and it chews through it.
I also bought the entire computer for $2500+tax new from Microcenter in December and it's appreciated 25% in value since then.
This only goes back one year, to Linux 6.18. Wins would be even bigger if we go back another year. Also, this isn't tracking any of the rest of the improvements in userland: it's just the kernel. Some newer GCC, and upcoming new x86-64v3 targets will all have some pretty nice wins too.
Great days to be on open source. And it only ever gets better.
Yes. AMD now has a complete portfolio of semiconductors: Ryzen laptop and desktop CPUS, Epyc data center CPUs, Instinct AI GPUs, Supercomputer CPU/GPU hybrid systems with FP64 support, Helios Rackscale AI, Gaming GPUs and consoles, Xilinx FPGAs.
Intel has a cache-as-ram (CAR) mode / Non-Eviction Mode, which is what you wanted but it is only available before DRAM is brought up via core initialization. Not sure about AMD.
AFAIK AMD doesn’t do this because it’s a pain in the butt to do from a hardware and firmware perspective. They basically have their equivalent of the intel ME do RAM for training and then the main cpu comes on with DRAM already up. At least it was like that five years ago when I last worked on an AMD part.
I'm not sure focusing on just 3 processors is enough to claim that Ryzen got 50% faster. There were already other, faster Ryzen processors available without the additional cache. The extra cache was a very new thing (and very temperature sensitive), so it makes sense that they start carefully and then get a lot of improvement quickly). And the 5800x3d was for an older chipset, with all the limitations that come with that.
It's not that temperature sensitive, 90C is plenty. If you can't keep the non-x3d processors underneath that you're just throwing away performance anyways
Are you? I'm pretty sure when I bought my 7800x3d, everybody was saying you had to be very careful about heat, because these damage more easily from high temperatures than regular CPUs. I just checked, and they do throttle at a lower temperature than others: 89 vs 95 degrees C.
For performance reasons you need to be careful with heat on the X3Ds because it will throttle quicker (partially because of the lower limit, partially because the extra cache sits between the CPU and the cooler making it harder to cool) but you don't have to worry about it from a damage perspective for the same reason: the limit is the point at which it'll drop performance to protect itself.
The x86 architecture still going strong. I've always though Ryzen was a fitting name, AMD had Ryzen from the ashes and struck a blow to the long dominant Intel.
When I built my Ryzen 1800X desktop, I named it "AMD Ryzing". My optimism turned out well placed.
I recently upgraded from that to a Core Ultra 7 that I got from work (e-waste recycling). It was not my intended upgrade path, and I hope Lisa isn't too mad (I kept my Radeon though).
Sort of: there were about half a dozen in micro PCs in a load picked up from a hospital. I was kinda bummed that they were BIOS password locked (as were the other PCs from there). The model didn't have a password reset jumper (or any reset mechanism outside of contact Dell support with proof of purchase), so I couldn't sell them as whole systems, and stripped them for parts.
We get DDR5-based systems from time to time, but they're understandably rare. Most of the time, they still work. My rule: any day I come across DDR5 is a good day.
I don't think you are correct for single-core performance.
A20 Pro scores around 4725 at 8.9w. Geekerwan's review showed that the iPhone 18 pro could dissipate as much as 6.4w during long gaming tests. Cutting power by 50% likely still keeps around 80% of the clockspeed.
As 9950x scores around 3400-3450 in geekbench, or around 28% slower.
There's a very good chance that the iPhone single thread performance is genuinely faster even during sustained loads.
Sustained performance is usually based on cooling, not the chip. There is no reason the iPhone chip can't sustain its performance if you give it a small fan.
So whatever your argument is, it doesn't make much sense.
The M6 is based on A20 Pro and it sustains its performance in a Mac Mini indefinitely - likely longer than Ryzens.
The 9800X3D doesn't pull anywhere near that much in single threaded workloads, it'd fry the core. It's also way less efficient than the iPhone CPU design and on an older and less efficient semiconductor process from TSMC.
If you compare like for like, ie. take the process node out of the picture, then you need to compare the Ryzen 9000 series to the A16, as both use TSMC's 4NP process.
It's impressive and yet it surely depends what you're doing with it? 50% better performance isn't going to make a local LLM feel quick and yet so many other aspects of computing are quite fast anyhow. I can browse the web fairly comfortably on a Raspberry Pi and it's a bit slow but manageable.
If your goal is web browsing you will want a faster single core performance. Unfortunately that means Apple M series. If your goal is local LLM, I’m afraid a years-old GPU will smoke the fastest CPU.
IIRC, compared to GPUs, M-series is currently still stuck in memory bandwidths from around 2016. M7 might catch up to 2019 or so. So GPUs will be better for LLMs for a pretty decent while.
How so? I don't know of any other mainstream platform with >1TB/s memory bandwidth. Personally I don't want to deal with macOS but between the memory bandwidth and out of box Thunderbolt networking it's hard to argue that Apple doesn't have a couple significant advantages over the current alternatives.
> I don't know of any other mainstream platform with >1TB/s memory bandwidth.
I mean, RTX 5080 has nearly 1TB/s, 5090 has nearly 2TB/s. Maybe you are talking about CPUs / unified memory platforms? I agree nobody else does it better. But for LLMs, GPUs can still be significantly faster than even the most advanced Apple silicon on the planet. TTFT in particular is super inferior with Apple, for now.
That's probably also the reason Apple had to reluctantly give into Nvidia servers for the initial rollout of Siri AI, though they claim to use trusted computing extensions to reach an acceptable level of privacy. (I do not trust that nearly as much as the Apple Silicon nodes)
It'll be amazing five years or whatever down the line to see Apple reaching those figures. They seem to be heading in that direction lately.
I suppose it depends what models, but even heavily quantized the mid tier local models need >100GB of memory so unified memory platforms are the only option I would consider a mainstream option (doesn't require special order etc). If you want to do it in two slots that's two Blackwell RTX 6000s which is around $50k list just for the cards, and if you want to do it in four+ slots that's outside the realm of normal desktop PCs. Right now 2x DGX Spark, 2x Ryzen 395, or 1x Mac Studio are the three configurations that get to 256GB of reasonably fast memory that you can order online and plug in, for around $8k to $12k depending on the details.
I Picked a threadripper for a box that needed a lot of IO and figured that with that many PCI lanes I couldn't go wrong. But I have to admit I've been more than pleasantly surprised by the performance of the CPU as well, it - easily - outperforms all of the XEON and I7 based boxes that I have. The only thing I wished I would have done different is to max it out with 256G DIMMs when they weren't the price of a car.
Zen 5 really is nice. I swapped my RAM from a Xeon W Sapphire Rapids machine into a Threadripper Pro 9000 series machine and I get almost double the memory read performance, plus it's a heck of a lot faster in single and multi core performance, and it runs cooler and quieter. Huge win all around. Aside from the price... (I went from Xeon W5-3435X to TR Pro 9985WX, eep.)
I have the complete opposite experience. Got a 9550x with 64gb ddr5 and a fairly high end mobo about two year ago. Just running the memory at stock speed. About 50% of the time I’d reboot and one or both of the sticks would only be detected as 2gb. Would need to do a hard shutdown to get it back.
I eventually gave up and turned off memory context restore and now I just deal with the minute plus (!) time to Post.
Not sure if amd memory controllers are just garbo or what but I’m going back to intel next chance I get.
Yea I’ve done endless googling around the issue and it seems to be a not uncommon issue with ddr5 but there’s no way I’m buying new parts now so I’ve come to peace with it.
On the desktop Intel has better memory controllers last time I looked, but it’s not that big a difference. Something in your setup is just broken. Failures will happen with any brand, so I wouldn’t chalk it up to AMD vs Intel… figure out where your problem is and get the faulty part replaced if you can.
I feel you though, regardless of the cause, that is an extremely frustrating place to be.
Yea I’m just bummed because it’s a real expensive time to be swapping memory parts around. Like I said, things seem to just work without memory context restore so I’m fine with just paying the cost when I reboot once a week or whatever. I feel like the fact that disabling MCR fixes it should narrow down the issue to some part, I’m just not sure which.
Did you set it to do memory training on every boot? That's a big stability help.
Should also run Memtest86+ a few times. With some stick swapping you can determine if you have a bad stick or slot.
I went the opposite way and went back to Intel after many many years on AMD. I snagged a Xeon 654 Granite Rapids ($850) for a new workstation geared towards local inference. I don’t need a ton of CPU cores and all Granite Rapids have 8 memory channels where as you need very high end Threadrippers (9975wx for $4000) for an equivalent due to their CCD design.
As a bonus Granite Rapids is super power efficient and runs much cooler than my previous TR and Ryzens. Intel seems to be doing good things again! My only minor complaint is that P2P doesn’t work on my multiple GPU setup because every PCIe5x16 lane has its own dedicated root to the CPU, but all that bandwidth is useful for MoE models.
You could get a PCIe switch backplane, stick your GPUs on there and then have them communicate locally through the switch using p2p. For some GPUs you may have to use a hacked driver (google p2p 3090 for instance).
C-Payne makes excellent stuff but it is very pricey and supply is more than a little spotty, I've used pretty much all of his stuff by now and I am a very satisfied customer if I can order what I need, which more often than not is not the case.
A good alternative is ADT, they make excellent boards, there is a 4 and a 5 slot PCIe expander with a 88096 on it that works extremely well. I have two of these connected to my rig with their own power supply and four GPUs in each (and another two in the main machine). It's not exactly a portable affair (to put it mildly). Note that these won't work in a standard PC case due to the slot spacing so you'll have to rig something for that yourself (I use 2020 + some custom 3D printed fixtures).
Yeah, I had to go high-end to get one with enough CCDs to make use of the memory channels. I'm curious to know what all-read bandwidth you see when you run intel mlc on your system. My Xeon W5 was doing ~180GB/s after a lot of tuning, which was very disappointing. The Threadripper does ~320GB/s. From what I've heard, Granite Rapids is supposed to have solved the bandwidth problems that ailed Sapphire Rapids, so your numbers ought to be similar.
Different kinds too? You never know when you might need a ball peen hammer or a soft faced hammer or a rubber or deadblow mallet or a framing hatchet. Why take the chance?
They bought the ram they needed when they needed it, and probably some extra too. Buying more than necessary is inherently a speculative action - you are betting on needing it in the future, and you are betting that it will be more expensive in the future. If it maintains its price it doesn't matter if you buy it now or later, if it gets cheaper or gets superseded by better alternatives then buying early is essentially a loss.
So in my view, this situation is almost exactly analogous to stock market speculation.
In the computer world, wait a few years and things get cheaper. That was broken by an exceptional event. Who would have thought you can't afford DDR4 in 2026?
Just like an investment though - you need to have the money to play. I wasn't buying a computer at the time because [snip a long list of other toys I bought instead].
My current plan is to stick with my existing computers for a couple more years and hope this blows over. Time will tell.
I'm aiming for my machines to be too heavy and hard to open ;)
I put 256G in it wondering if I would ever need that much RAM and now it looks way too small. Highly frustrating. I wonder if the companies that made these deals realize that they've just given an entire industry a reason to look for alternatives. It's not like they wouldn't have sold their product otherwise.
I picked up a pre-ai-price-insanity AX162-R at Hetzner a while back and loaded it up on memory to max out the 12 channels the 48c EPYC 9454P as a "this will be the last mysql box I'll need" and have I been _wildly_ impressed with it's performance. The things I throw at it are honestly laughable at times, wildly irresponsible queries against a rather large database, the redis qps metrics are ridiculous and I just load it up with random ggufs since the memory bandwidth is... not terrible and 384GB of it is... useful.
The web app that's hosted on it deals with lots of images and text - over 100mil of each deduplicated, embedded, simhashed - it does the hashing, and the embedding in real time during ingest. Just handles it.
Weird they didn't talk about the DDR4->DDR5 transition. The 5800X3d->7800X3D went from DDR4->DDR5.
DDR5 is literally twice the price per GB. It's better for sure, higher bandwidth but waaaay more expensive. The 9800X3D supports much faster and even more expensive DDR5 than the 7800X3D too (DDR5-2667 to DDR5-5600).
It's funny how my 3950X, which I still find to be an absolute monster for anything I throw at it, isn't even in the benchmarks anymore. Hard to believe it's 7 years old.
Perhaps better looking at it without the 3D Cache on GB6 scores.
2020 Zen 3 AMD Ryzen 7 5800 - 1852
2024 Zen 5 AMD Ryzen 9600X. - 2914
2027? Zen 6 - Est 3200?
And compare this to Apple.
M6 4000+
M5 3600+
M4 3300+
Even the Snapdragon X2 Elite X2E is better at 3300.
The gap isn't exactly shrinking. And I would love x86 to prove me wrong. But even a hypothetical Zen 7 in 2028 ( 2029 for consumer ) may only catch up to Apple's M5 released in 2025.
So far, AARCH64's advantages mostly seem to affect single core performance. AMD is also behind in the nanometer race, which is one of the ways Apple managed to improve their peformance over the years as well. If they manage to catch up in terms of die shrink, I think their multi-core perf will once again put AMD in the lead for data processing at the very least.
Although, if you count EPYC, which you might as well with the prices Apple charges for their top-end models, AMD still beats Apple with ease at multi-processing.
It's a real shame that these processors are made by Apple (and Qualcomm, I suppose). They'd be my go-to chip if they weren't.
The AMD EPYC Zen 6 is built on TSMC N2, so it is a similar comparison including clock speed, x86 just hasn't delivered.
Apple has been iterating on a yearly basis, A17, A18, A19 and A20 are all different core design. Some of them just so happens to come with new node. On the other hand Zen 5 to Zen 6 took 24 to 30 months.
Huh no, wide vector units or lots of execution units per core is nothing like wide instruction words? Superscalar out-of-order execution is specifically a way to get instruction level parallelism without the wide instruction words
I was thinking that for this new generation, "Superscalar" doesn't cover it anymore, except as a generic term for the design technique. "Ultrascalar"? "Superscalar Pro"?
I mean superscalar just is a different name for using wide cores and runtime dependency tracking and out-of-order execution for instruction level parallelism without VLIW.
To my eyes, your message looks kinda like, "The term 'bicycle' doesn't cover it anymore, things have changed so much since they were introduced. We should call them something else, maybe 'bicycle pro'? 'super bicycle'?" No it's just a bike, and no, modern CPUs are just CPUs which look scalar at the ISA level but have ILP
Sure, but in my mind it feels like the CPU is slurping up individual instructions until it has enough to execute a whole bunch of them at once in one step (which would be equivalent to a big VLIW).
It's not really like that, it's more that it slurps up individual instructions, executes some of them just in the front-end via register renaming and dispatches the rest to available execution units, then retires them as the execution units are done. Sometimes it fuses together a couple of instruction (macro-op fusion) but my understanding is that that's much more limited, like fusing "compare then conditionally branch" sequences into single compare-and-branch μops
191 comments
[ 0.28 ms ] story [ 120 ms ] threadBetter thermals maybe? Less throttling?
One of the biggest improvements to the Ryzen 7 X3D series (e.g. 9800) is that the caches have been moved from one side of the die (to the other), which places the major heat source closer to the heat sinks.
SO yes, less heat throttling.
Power is heat, and the amount dissipated is the square of the voltage- a processor that is reliable at less voltage means you can get a lot more frequency in the same heat envelope.
Yet reliability decreases as frequency goes up, and that can only be stabilized by adding more voltage- so the faster you run the processor, the more voltage you ultimately have to give it, so the power/heat produced grows exponentially until you can't get rid of the heat fast enough (at which point your only option is to actively cool the chip).
This is overclocking 101.
Note that classic overclocking was viable because of arbitrage- buying a processor, pushing it to the point it got too hot, and stress-testing it at that temperature to ensure reliable operation. Processor manufacturers all do that formal verification at the factory now as they'd be uncompetitive otherwise, especially in laptops.
First zen 3 November 5, 2020 with desktop processors. First desktop Ryzen 9000 processors on August 8, 2024. So the generations were about 4 years apart.
>Compared this to the article "How did AMD Ryzen get 50% faster in two years?" The Zen 3 uArch being used in the article came out in 2020. So it is more like AMD got 50% faster in 5 years.
[1] https://news.ycombinator.com/item?id=49772502
Not implying anything here, but my comment was actually before yours, this article was reposted via HNs 2nd chance pool, https://news.ycombinator.com/pool
From Geekerwan and many other Chinese sources. The M6 CPU Core is more like an M5 die shrink.
That's pretty much irrelevant since the AMD's FX (Bulldozer and Piledriver) arch's issues weren't that SW at the time wasn't using all the 8 cores. Intel dropped the Core 2 Duo and Quad into the era where most SW was still single threaded and those CPUs still ripped regardless.
Here's the big reasons why the FX sucked back then and why they still sucks today in the multicore SW era:
So these are the main bottlenecks today regardless of parallelization of desktop SW.AMD was also having to deal with the fact GloFo split off and was relying more on general 'bulk' lithography, which kneecapped them for some time especially due to yield issues on the FX series and overall cost of that deal.
Intel also very quickly after, released Sandy Bridge and aggressively scaled it up and down; the 2500K was so cheap yet powerful I know of at least one setup that ran for a decade an only got replaced because they needed to upgrade to windows 11 for compliance-esque reasons. My own 2500K I replaced in 2017-2018-ish, only because either the motherboard took an unfortunate dive and it was easier to replace both at once.
FWIW, I did do a cheapie FX build in 2015ish for my then-girlfriend as a DVR and light gaming/emulation 'under the TV box', and it did the job well for the price, but it definitely wasn't anything amazing.
It was a tough time for AMD for sure. I think the 'split' between the Cat cores (Bobcat/Jaguar) also hurt them from a resource standpoint, although one could argue that it also kept them alive to recover (i.e. Jaguar in XBox One and PS4 being a volume contract part) [0]. They did a lot of moves that caused short term pain (that glofo spinnoff helped pay off the ATI Acquisition AFAIR) but helped them become the company that is still surviving today.
[0] - One odd side note, I still find it odd that they never did a dual channel Jaguar laptop part. I still ask whether it was because it would have made the FX look that bad...
Yeah that's hyper-threading intel was doing it as well and all modern CPUs do it as well. Where AMD dropped the ball, was they did not disclose that in their marketing as clearly as they should.
All CPUs today are marketed as x cores 2x threads, back then some AMD marketing genius in their infinite wisdom put 8 cores on the box, instead of the honest 4 cores with hyperthreading.
They did just fine in parallel workloads, so I think this is not accurate. The design scaled just fine. The problem was that each core was weak.
They in fact DID NOT "do just fine in parallel workloads". The contemporary competition from Intel that was 4-core + SMT was beating AMD's 8-core FX CPUs in plenty of multi core tasks. AMD only won in >4-thread strictly integer benchmarks.
The myth that each two-core module functioned more like one core with hyperthreading would suggest that these CPUs would have much higher per-core performance when lightly loaded than when fully loaded. That is not what happened. Each core was crummy even when lightly loaded, but under full load you would have eight crummy cores, which would beat four Intel cores on a lot of workloads.
The only time contention was a serious problem was with workloads that were dominated by floating point, which were relatively rare.
By that definition it definitely was not scalable.
>But we did in fact see a roughly 8x speedup with eight threads,
Care to share a source? Because AFAIR there definitely was no 8x linear speedup with 8 threads even in benchmarks, let alone in real world use cases. The only benchmarks where those 8 threads would scale best and beat Intel were archival compression/decompression and media encoding. At everything else Intel wiped the floor with it.
>The only time contention was a serious problem was with workloads that were dominated by floating point, which were relatively rare.
Many real-world compute workloads, especially gaming related, are floating point.
This argument has never made very much sense to me. Yes, the decision to group cores into modules with some shared resources did introduce a bottleneck and result in lower performance than having isolated cores. But this didn't make the processor worse than if it didn't have the additional cores at all. These processors were at their best on highly parallel workloads, despite the presence of this bottleneck. AMD shipped far more cores than Intel at the same price and as a consequence came out ahead on highly parallel workloads. The contention for the shared front end was not serious enough to make up for the core count advantage. How the module architecture an explanation for why the design failed, when the bottleneck only becomes relevant in situations where the design is winning?
It's hard to find benchmarks from 15 years ago. But Phoronix finds what I describe: the design doesn't scale quite as well as a "true" eight-core design, but still scales better than the competition of the time which had lower core counts. Though it does seem straightforwardly bad at parallel floating point workloads.
https://www.phoronix.com/review/amd_bulldozer_scaling/7
It's as FlowingRiver said. If we were to release the same parts again with modern software, Bulldozer would be a far stronger competitor to Sandy Bridge. FX aged better than Intel's designs of the same era. But Intel's designs were better to start with, so I'd still take the Core.
The reason they sucked is that they ran hot and their per-core performance was terrible.
So in a physical heavy computation task, involving double precision math (where the AMD design of share FPU units should penalize most), the FX cores where happy churning numbers with the expected speedup Vs a single core/thread version of the code. So stop saying that FX cores sucks at multithread. They fucking worked fine on that kind of tasks.
Steamroller did see 30% IPC improvements over Bulldozer (all the design work would have been done before they switched to Zen), but AMD canceled the full FX version, and only ever shipped the APU version of Steamroller (with only 2 modules, aka 4 threads).
If they had shipped a Steamroller FX cpu, the generational improvements would have looked similar to many of the generational improvements that Zen received... but didn't really matter as Bulldozer started so far behind.
My hope is on Huawei, which has recently developed a stacked chip architecture that reminds me of HBM, and I'm wondering if that's step 1 of China entering HBM production.
I don't know enough about hardware, so I might be saying absolute nonsense, happy to be corrected so I can learn where I'm wrong
It has the highest base clock of any modern AMD chip at 4.3 GHz, plus 16-cores, which is already more than most workloads need.
Can’t wait to see what Zen 6 brings early next year. Rumors say 24-cores and another big jump in single-core performance.
All stated for gaming, that is.
It’s even difficult to find CPU benchmarks that don’t overemphasize 1080p and eSports scenarios.
My upgrade from 5600x3D to 9850x3D felt kind of dumb at the time, but I decided to do it because Micro Center’s bundle deals are so far below market pricing.
I was shocked at how much better it is. Benchmarks and FPS don’t really show things like micro stutters and little performance wrinkles like that. I’m not even sure 1% low FPS counts capture it.
A great example game for this is Oblivion Remastered. Upgrading my CPU alone with the same GPU took away the environment loading slowdown almost entirely.
The benchmark will tell you that I didn’t gain any FPS during gameplay but every time I open a door into the new environment my CPU is positively impacting the experience.
I would have said the exact same thing you are saying until I actually experienced upgrading to the best on the market. For the record, this is the first time in my life I’ve actually owned the best CPU on the market for gaming.
My old advice would have been to buy one of those sweet spot cheaper mid-range gaming CPUs, but my newer advice is really if you’ve already spent all that you’re willing to spend on a GPU (I have a 9070XT, my only upgrade paths are insanely expensive), buy the highest gaming CPU on the list for gaming benchmarks (e.g., I wouldn’t go crazy with a 9950X3D2 since it doesn’t have any gaming improvements above the 9850X3D).
And the thing about CPUs is they’re not insanely expensive like GPUs. We are talking a price delta of $200 between this beast of a CPU and something way more middling.
Given the massive increase in hardware prices lately, we are going to have to get by with mid range or even low end hardware for a lot longer. Studios will be forced to actually optimise games to not run like shit on a sub $8000 PC.
Every game should be targeting the switch 2 and steam deck in terms of power.
I think PS5 is a perfectly reasonable spec to be targeting, which was the higher end of mid range about 5 years ago, and still significantly out performs Switch 2 (which significantly out performs steam deck). I’m sure the GTA sales will demonstrate that the install base at this spec is wide enough for broad commercial success.
I grew up with gaming systems that got rapidly better and while it can be argued that we are at a diminishing returns plateau of gaming horsepower, I don’t think the solution is to never get hardware upgrades ever again, even in the current RAM-constrained environment.
We saw with much of the PS5’s lifespan that many of its cross-platform games came out on PS4 and original Switch with minimal compromise.
Not enough of the install base owns those higher end solutions, so for the most part someone with a better performing rig is getting better FPS, higher resolution, and not much else. Even gaining additional draw distance is rare these days, and higher resolution textures are basically impossible to see.
If the console is going to survive rather than decline as a whole I think that Microsoft and Sony will need to get more serious about innovation rather than just making a samey spec bump at a high price that doesn’t motivate their users to upgrade. Nintendo already proved that basic and rather mild level of innovation can excite the market (original Switch).
It's relatively easy to allow powerful hardware to just run the game at 4k 120fps highest settings while weaker hardware runs exactly the same game at lower graphic settings.
I think it's limited by their software. The hardware buys them flexibility. Nobody is pushing it like the ps2 days.
Having a powerful rig is the brute force way around the game’s performance problems.
My PC wasn’t $8000. My 9850X3D, 32GB of RAM, and motherboard cost $800 in a bundle. Yes, purchased in 2026 during the RAM crisis. I was also able to sell my previous CPU, RAM, and board for around $500.
Obviously this is not a budget build but about half the cost is the graphics card. I would have saved minimal amount of money if I had gotten a worse CPU.
Does the base clock even mean anything beyond something that may or may not vaguely indicate the performance? Its lower than what these chips can sustain with sufficient cooling from my experience. It's higher than what the chips will clock down to when not sufficiently cooled.
They aren't the maximum or minimum but I think they are the only clocks with a specific guarantee "if you match X, you must be able to get Y". The max boost clocks usually aren't too hard to reach if you're trying though, there is just no guarantee you will in all workloads or how much it takes to get there.
Let's not discuss battery life and overall size, however it helps the thermals.
I do a lot of CPU bound engineering and scientific compute on the computer and it chews through it.
I also bought the entire computer for $2500+tax new from Microcenter in December and it's appreciated 25% in value since then.
This only goes back one year, to Linux 6.18. Wins would be even bigger if we go back another year. Also, this isn't tracking any of the rest of the improvements in userland: it's just the kernel. Some newer GCC, and upcoming new x86-64v3 targets will all have some pretty nice wins too.
Great days to be on open source. And it only ever gets better.
I recently upgraded from that to a Core Ultra 7 that I got from work (e-waste recycling). It was not my intended upgrade path, and I hope Lisa isn't too mad (I kept my Radeon though).
How the hell does a high end CPU that's barely 2 years old end up in e-waste recycling? Was Richie Rich using them or something?
Sort of: there were about half a dozen in micro PCs in a load picked up from a hospital. I was kinda bummed that they were BIOS password locked (as were the other PCs from there). The model didn't have a password reset jumper (or any reset mechanism outside of contact Dell support with proof of purchase), so I couldn't sell them as whole systems, and stripped them for parts.
We get DDR5-based systems from time to time, but they're understandably rare. Most of the time, they still work. My rule: any day I come across DDR5 is a good day.
The x86 folks lost the ball completely. For workstation/ mobile needs you go to arm. For matmul at scale you go to gpus.
I guess windows gaming with separate gpu is the only remaining market for them.
Unless Nvidia releases a motherboard with an arm cpu on top of their RTX cards, and then that market is gone too.
There is no chance the sustained single core perf is better than a 150=200+w ryzen with a big noctua cooler on it.
A20 Pro scores around 4725 at 8.9w. Geekerwan's review showed that the iPhone 18 pro could dissipate as much as 6.4w during long gaming tests. Cutting power by 50% likely still keeps around 80% of the clockspeed.
As 9950x scores around 3400-3450 in geekbench, or around 28% slower.
There's a very good chance that the iPhone single thread performance is genuinely faster even during sustained loads.
So whatever your argument is, it doesn't make much sense.
The M6 is based on A20 Pro and it sustains its performance in a Mac Mini indefinitely - likely longer than Ryzens.
>Single core performance still worse than an iPhone.
I can believe the core with active cooling can sustain good performance at least, but not in the IPhone.
I mean, RTX 5080 has nearly 1TB/s, 5090 has nearly 2TB/s. Maybe you are talking about CPUs / unified memory platforms? I agree nobody else does it better. But for LLMs, GPUs can still be significantly faster than even the most advanced Apple silicon on the planet. TTFT in particular is super inferior with Apple, for now.
That's probably also the reason Apple had to reluctantly give into Nvidia servers for the initial rollout of Siri AI, though they claim to use trusted computing extensions to reach an acceptable level of privacy. (I do not trust that nearly as much as the Apple Silicon nodes)
It'll be amazing five years or whatever down the line to see Apple reaching those figures. They seem to be heading in that direction lately.
I eventually gave up and turned off memory context restore and now I just deal with the minute plus (!) time to Post.
Not sure if amd memory controllers are just garbo or what but I’m going back to intel next chance I get.
Up to date on bios updates?
I feel you though, regardless of the cause, that is an extremely frustrating place to be.
It is, but isn’t it all warranty in your scenario?
It would be painful to claim though, as what part is at fault?
As a bonus Granite Rapids is super power efficient and runs much cooler than my previous TR and Ryzens. Intel seems to be doing good things again! My only minor complaint is that P2P doesn’t work on my multiple GPU setup because every PCIe5x16 lane has its own dedicated root to the CPU, but all that bandwidth is useful for MoE models.
A good alternative is ADT, they make excellent boards, there is a 4 and a 5 slot PCIe expander with a 88096 on it that works extremely well. I have two of these connected to my rig with their own power supply and four GPUs in each (and another two in the main machine). It's not exactly a portable affair (to put it mildly). Note that these won't work in a standard PC case due to the slot spacing so you'll have to rig something for that yourself (I use 2020 + some custom 3D printed fixtures).
Not sure I’ll need to upgrade my computer ever again :D
One is a tool, the other is a speculative financial asset.
So in my view, this situation is almost exactly analogous to stock market speculation.
My current plan is to stick with my existing computers for a couple more years and hope this blows over. Time will tell.
They (in theory) could padlock them so the drives and ram couldn't be stolen?
Wikipedia describes but lacks pictures.. https://en.wikipedia.org/wiki/Computer_lock
I put 256G in it wondering if I would ever need that much RAM and now it looks way too small. Highly frustrating. I wonder if the companies that made these deals realize that they've just given an entire industry a reason to look for alternatives. It's not like they wouldn't have sold their product otherwise.
The web app that's hosted on it deals with lots of images and text - over 100mil of each deduplicated, embedded, simhashed - it does the hashing, and the embedding in real time during ingest. Just handles it.
These things are absolutely insane.
DDR5 is literally twice the price per GB. It's better for sure, higher bandwidth but waaaay more expensive. The 9800X3D supports much faster and even more expensive DDR5 than the 7800X3D too (DDR5-2667 to DDR5-5600).
Even really good DDR5 could be had for cheap; I remember paying like 300$ for 96GB of high-quality M-die hynix DDR5.
Zen 3: 2020 release
Zen 4: 2022 release
Zen 5: 2024 release
The author is disingenuous by looking at only the X3D variants. The Zen 3 X3D variant came out late.
For comparison, in the same time frame, the M1 went is 52% faster in ST and 72% faster in MT.[0]
[0]https://browser.geekbench.com/v7/cpu/compare/433673?baseline...
2020 Zen 3 AMD Ryzen 7 5800 - 1852
2024 Zen 5 AMD Ryzen 9600X. - 2914
2027? Zen 6 - Est 3200?
And compare this to Apple.
M6 4000+
M5 3600+
M4 3300+
Even the Snapdragon X2 Elite X2E is better at 3300.
The gap isn't exactly shrinking. And I would love x86 to prove me wrong. But even a hypothetical Zen 7 in 2028 ( 2029 for consumer ) may only catch up to Apple's M5 released in 2025.
Although, if you count EPYC, which you might as well with the prices Apple charges for their top-end models, AMD still beats Apple with ease at multi-processing.
It's a real shame that these processors are made by Apple (and Qualcomm, I suppose). They'd be my go-to chip if they weren't.
Apple has been iterating on a yearly basis, A17, A18, A19 and A20 are all different core design. Some of them just so happens to come with new node. On the other hand Zen 5 to Zen 6 took 24 to 30 months.
To my eyes, your message looks kinda like, "The term 'bicycle' doesn't cover it anymore, things have changed so much since they were introduced. We should call them something else, maybe 'bicycle pro'? 'super bicycle'?" No it's just a bike, and no, modern CPUs are just CPUs which look scalar at the ISA level but have ILP