Ryzen 7000 is launching later this month, and we're starting to see benchmarks of official Ryzen 7000 chips in the wild. But with the introduction of AVX512 on Zen 4, the new Geekbench benchmarks inflate Ryzen 7000's average performance estimates over Zen 3 chips to seemingly absurd levels.
@Benchleaks on Twitter shared a new Ryzen 9 7950X Geekbench 5 score that shows very impressive gains for the Zen 4 architecture over Zen 3.
The test bench for this new Geekbench 5 result featured a 16-core Ryzen 9 7950X Zen 4 chip with a maximum reported frequency of 5.738GHz, 32GB of 6000MT/s DDR5 RAM, and an Asus ROG Crosshair X670E Extreme motherboard.
The test results show a single-core score of 2,217 points, and an eye-watering multi-core score of 24,396 points.
For comparison, we pulled a recent Geekbench 5 benchmark of the 16-core Ryzen 9 5950X Zen 3 processor (running on a B550 Aorus Elite motherboard with 32GB of 3600MT/s DDR4 memory). The 5950X's test results show a lower single-core score of 1,725 points and a dramatically lower multi-core score of 17,069 points. Based on these results, the 7950X outperforms the 5950X by 29% in the single-core test and 43% in the multi-core test.
These scores look fantastic for the 7950X, but they don't tell us the entire story. For that, we need to look at the different workloads Geekbench 5 uses in its benchmark. Geekbench 5 splits three types of workloads across 21 benchmarks, including integer, floating point, and AVX-intensive cryptographic workloads.
If we take a look at how both chips perform in each of these different workload categories, the story becomes clearer.
|Workloads - Single Threaded||Ryzen 9 5950X||Ryzen 9 7950X||Row 0 - Cell 3|
|Workloads - Multi Threaded||Row 4 - Cell 1||Row 4 - Cell 2||Row 4 - Cell 3|
While the floating point and integer results are impressive on their own, Ryzen 7000's biggest performance gains come from the chip's substantially higher Cryptograph benchmark results — which are almost 71% higher than the 5950X's results in single-core performance.
This unexpected leap in performance can be attributed to Ryzen 7000 adding the AVX-512 instruction set, which Ryzen 5000 lacks. AVX-512 is one of the newer instruction sets seen on modern processors, and it can have a remarkable performance uplift in apps that support it.
The problem with AVX-512 is its adoption rate, which is very low by today's standards. Despite being available for over five years, very few apps currently leverage it — as a result, only a minority of power users and content creators are able to use AVX-512's capabilities.
By contrast, integer and floating-point workloads are the most common workloads you'll see on processors today. Gaming, multitasking, production, and just about everything else uses some form of integer instructions or floating-point calculations.
In other words, Ryzen 7000's Geekbench 5 performance results could be considered somewhat misleading. The crypto scores, while incredible to see, will only really affect a small fraction of Ryzen 7000 users. On top of this, Geekbench 5 has been known to weight its average scores to give some subtests more importance than others.
Unfortunately, this has been an ongoing issue with some subtests — we saw this exact issue with the Core i9-11900's Geekbench results. But with the introduction of AVX-512 on newer processors — including Intel's 11th Gen Rocket Lake platform, these wide performance deviations between workloads have become much more prominent — making general performance estimations difficult with some synthetic benchmarks.
That said, Ryzen 7000's real-world performance is very good in Geekbench 5, with integer and floating-point scores well ahead of Ryzen 5000's. But the addition of AVX-512 combined with AVX-512-intensive benchmark runs make Ryzen 7000's single-core performance look just inflated like we saw with Rocket Lake CPUs — it's good, but it's not that good.
You have a fair technical point re @Benchleaks results, but if read quickly your article comes across as challenging all the claims so far.
Yes, minus Alder Lake which doesn't enable AVX512 by default and is fully disabled on the latest versions.
It's important to note that the Geekbench results don't conform to the performance increase users will see right away, but it should be noted the situation is likely to improve; otherwise, it would be like saying of the original Ryzens that having 8 cores is meaningless, because real software doesn't try to use more than four. That was true, but it changed quickly.
Of course, Intel chips are likely to start having AVX-512 again soon, now, and their implementations may be superior to AMD's. (Isn't competition wonderful?)
It's in the article below regarding AVX on Rocket Lake.
Maybe you should just delete that post.
"In a nutshell, you shouldn't trust Geekbench 5's overall scores as an accurate measure of Rocket Lake's performance, and there's a technical reason why. We've encountered strange phenomenons with Geekbench 5, where its use of AVX-512 can widely skew the results in the encryption subtest. "
They put it in a small selection of mostly server CPUs, so it was hard to get practice writing code.
Each generation supports a subtly different instruction set.
The entire CPU slows down if any AVX-512 instructions are being executed.
I hope AMD standardizes on a single version across all Zen 4s, and ideally it runs without slowing other programs. Then AVX-512 can get more wide spread acceptance.
Now we know why. Apparently the desire to bury Intel intruded well into the land of make believe?
Shame, the story was already good but from the start you could easily see the push-pull between the "honest" 15% alleged performance gains and the "complete boolsheet" 40% that some of the marketing folks wanted to push.