Wednesday, May 29th 2024
AMD Ryzen 9000 Zen 5 Single Thread Performance at 5.80 GHz Found 19% Over Zen 4
An AMD Ryzen 9000 "Granite Ridge" desktop processor engineering sample with a maximum boost frequency of 5.80 GHz was found to offer an astonishing 19% higher single-threaded performance increase over an AMD Ryzen 9 7950X. "Granite Ridge" is codename for the Socket AM5 desktop processor family that implements the new "Zen 5" CPU microarchitecture. The unnamed "Granite Ridge" processor comes with an OPN code of 100-0000001290. Its CPU core count is irrelevant, as the single-threaded performance is in question here. The processor boosts up to 5.80 GHz, which means the core handling the single-threaded benchmark workload is achieving this speed. This speed is 100 MHz higher than the 5.70 GHz that the Ryzen 9 7950X processor based on the "Zen 4" architecture, boosts up to.
The single-threaded benchmark in question is the CPU-Z Bench. The mostly blurred out CPU-Z screenshot that reveals the OPN also mentions a processor TDP of 170 W, which means this engineering sample chip is either 12-core or 16-core. The chip posts a CPU-Z Bench single-thread score of 910 points, which matches that of the Intel Core i9-14900K with its 908 points. You've to understand that the i9-14900K boosts one of its P-cores to 6.00 GHz, to yield the 908 points that's part CPU-Z's reference scores. So straight off the bat, we see that "Zen 5" has a higher IPC than the "Raptor Cove" P-core powering the i9-14900K. Its gaming performance might end up higher than the Ryzen 7000 X3D family.
Many Thanks to TumbleGeorge for the tip.
Source:
Wccftech
The single-threaded benchmark in question is the CPU-Z Bench. The mostly blurred out CPU-Z screenshot that reveals the OPN also mentions a processor TDP of 170 W, which means this engineering sample chip is either 12-core or 16-core. The chip posts a CPU-Z Bench single-thread score of 910 points, which matches that of the Intel Core i9-14900K with its 908 points. You've to understand that the i9-14900K boosts one of its P-cores to 6.00 GHz, to yield the 908 points that's part CPU-Z's reference scores. So straight off the bat, we see that "Zen 5" has a higher IPC than the "Raptor Cove" P-core powering the i9-14900K. Its gaming performance might end up higher than the Ryzen 7000 X3D family.
132 Comments on AMD Ryzen 9000 Zen 5 Single Thread Performance at 5.80 GHz Found 19% Over Zen 4
As mentioned above, the IPC increase from Zen 3 to Zen 4 was a miserable 1%, and CPU-Z obviously doesn't recognise (or is "optimised" to be running slow on) the AMD Ryzen CPUs properly, so we can't really expect over 18% IPC increase, unless the bench is changed. I don't know how the newer 19.01.64 AVX2 (beta) and 19.01.64 (beta) compare to the default 17.01.64 test.
Anyway the point is that 900 1T cpu-z score for next gen AMD is bad since it hardly matches Intels curren't gen. And let's not even mention Apple M4 which blows them all out of the water.
Extremely conservative generation...
I'll be buying a ZEN 5 3D (for Gaming) so I hope they improved the latency with the interconnect bandwidth. 19% more performance for a new architecture is not bad at all, but it's not crazy good either! ZEN 5 was hyped as the Messiah of CPUs so yeah. Maybe ZEN 6 will be the true "Game Changer"
All Zen CPUs (that is: Zen 1, Zen 2, Zen 3, Zen 4) had 6-wide dispatch since the beginning, while at the same time the IPC difference between Zen1 and Zen4 is fairly large, thus if by "Repipelined frontend" AMD in their own Zen5 slide meant changes in the frontend to supply the 8-wide dispatch in Zen5 with enough µops per clock cycle, then the claim about "Repipelined frontend" would be quite accurate and would reflect reality. Whether the term "New grounds up microarchitecture" used in AMD's slide was an overly bold statement or wasn't is to be seen about 1 week from today - so we wait.
Well, of course AMD is *today* busy finishing the Zen6 design (in parallel to launching Zen5) and already knows the approximate/projected/simulated IPC of the future Zen6 CPU. But we need to wait - so we wait. Of course we are encouraging them to their best and push some limits!
From git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/patch/?id=29c73fc794c83505066ee6db893b2a83ac5fac63 (file pipeline.json):
But I will say that IMO Zen 4 and AM5 in general is very disappointing. AMD did not go nearly far enough and left themselves behind Intel until the x3D chips came out. Zen 5 really needs to address this and make the AM5 platform worth the high cost of membership. I hear Zen 5 will bring with it a refresh of AM5 with USB4 support via a MediaTek chip, which I really hope doesn't suck. Memory support on AM5 is a joke, but I hear the refresh may tackle this with an IF of 2.4GHz, officially supporting DDR5 8000, AMD desperately need this, as their memory controller on Zen 4 really is a complete joke compared to what Intel gets, even at the same speed. I also hope they also sort out the PCiE5 issues and make it more widespread than just the most expensive motherboards. But it's AMD, and they have never been particularly great with chipsets.
On a particular workload I run, the difference in performance between EXPO 6000/2000 Mem/IF and Tuned 6000/2033 Memory/IF is 11% on the 7950x3D.
I hear that because AMD would not use TSMC's newer 3nm class, they stuck with a slightly improved 4nm node, meaning that they have not been able to do anything more radical for Zen5. They have also kind of designed themselves into a bit of a corner regarding their chiplet design and its physical size. The AM5 socket is simply too small to be able to fit more chiplets. They needed to reduce the physical size of the chiplet before they can go and put more cores on it. They can't do anything about the IOD due to it not scaling with smaller process nodes, so only expect that to get physically larger as they are forced to add more features to it.
I hear that the new Zen5 chips, when paired with a "refreshed" AM5 motherboard will officially support DDR5 8000. But it is not clear if this is because the IF is clocked at 2.4GHz or they have simply bolted on yet another memory divider, which will only make their memory situation worse. Hopefully we will find out soon. But that leaves a bad taste in the mouth that you might need to go out and spend $300 - $500 on a "new" AM5 motherboard, just to correct something that should have been a day one feature of the AM5 platform.
This is the efficiency graph. Yes, the x3d looks horrible in ST efficiency, it loses in efficiency by ~50% compared to the similarly performing 13400f. I don't know how AMD achieved that amount of inefficiency, but oh boy.
Now let's see you comment on intel's efficiency. I mean, if you have no idea how AMD acheived that level of inefficiency, what words do you have for intel? Surely worse right?
Your MT graph is useless - yes obviously the low power zen 4 parts will be at the top cause there aren't any low power Intel parts on the graph. Put all the T and non k parts in that graph and you won't find a single AMD cpu on the top 10, lol.
And uh..no my MT graph isn't useless at all because there are literally like for like AMD vs Intel CPU's in there so it's an apt comparison. As you can see, Intel gets destroyed in there at every level. Your point about bringing in low power intel CPU's in the mix is funny. Why not compare directly comparable desktop CPU's like we are here? Not to mention, there are low power AMD CPU's too.
Your power consumption argument is so strange. It's like..Intel universally gets destroyed in power consumption from every reputable reviewer out there and suddenly you're like 'ohhhhh no no that's not the case look at a useless benchmark, intel uses less power here so it must be better'. PfftI should've just ignored his posts, I honestly forgot for a second he's around the same level as userbenchmark and that's saying something.
At ISO power limits most Intel CPUs are more efficient than their AMD counterparts in MT. By a lot.