Are These AMD Ryzen “Leaked” Benchmarks Legitimate?
During AMD’s New Horizon event, the company discussed their new processor range and inspiration behind the name. The upcoming Ryzen CPU was benchmarked against the i7-6900K which almost costs $1000. Despite having a slower core clock, AMD’s new CPU either matched the Intel equivalent or performed better while having a 95-watts TDP compared to the i7-6900K’s 140-watts figure. WCCFTech recently posted some questionable leaked benchmarks ahead of the product’s launch which makes for some interesting reading. Apparently, the new Ryzen CPU outputs a multi-threaded score of 1188 cb in Cinebench R15. To put this into perspective, the Intel Core i7-6900K tends to achieve a score around 1500 cb and the i7-6950X reaches 1800 cb. On another note, the i7-7700K at stock settings obtained a 966 cb score and 1083 cb when overclocked to 5GHz. Therefore, it seems the CPU is tailored more towards the i7-7700K’s performance area.
The Ryzen processor was also benchmarked in the Fritz Chess application and managed a 36.86 relative score. Interestingly, the i7-6950X achieved 51.50 and the i7-6900K reported a 47.80 result. Furthermore, the i7-7700K at the default Turbo frequency posted a 35.52 score and 41.44 once overclocked to 5.0GHz. The Fritz Chess Kilo Nodes Per Second is another important reading which provides an insight into the performance of various processors. The Ryzen chip obtained a 17693 score while the i7-6950X hovered around 24000 and the i7-6900K gravitated towards 22500. Next up is the i7-7700K which fell slightly behind the Ryzen CPU with a score of 17049 although the performance improved considerably at 5.0GHz and resulted in 19891.
To be perfectly clear, these benchmarks are unverified so it’s important to take the results with a pinch of salt.
UPDATE: This is not a source I would trust, because the Cinebench score is cropped and there’s no CPU-Z screenshot. I wanted to share this piece of information with our readers in case it’s of interest but the feedback has been useful and I’m very sceptical of the data acquired by WCCFTech. Any comments would be appreciated regarding this matter to help drive future content.
Normally you’d expect a screenshot with CPU-Z on the side to verify what is being screenshotted.
Yes seeing as it’s open and to the left of the Chess picture…
that is right
At the event last week, they also claimed that all Ryzen chips would ship at 3.4Ghz or higher, and demoed at 3.4Ghz with turbo disabled.
This article just says a “Ryzen CPU”. No indication of clocks, no indication of cores, turbo, model, engineering sample vs production variant, ect ect..
It can be fun to publish and speculate on unverified benches, but they are kinda useless without some accompanying system data.
Chessbase found 16 threads and that’s what was used. Still very sketchy.
It could be any 16 thread cpu really since no other info was give. For all we know it’s some Xeon with low clocks.
Fair comment, I’ve updated the article to make the benchmark appear less as “fact”.
At the Horizon event, AMD demoed a Ryzen chip vs a BDW-E chip with turbo disabled and dual channel, on a pair of cherry picked benchmarks. It seems people forgot how well Bulldozer performed on Blender and Handbrake. Blender and Handbrake are GPU-like benches. Also during the Horizon event, the 95W Zen chip was consuming about the same power than the 140W chip: 191W vs 187W; this is a 2% difference.
I don’t know if those recent leaks are legit or not, but I know that the scores agree with all the former leaks (AoS, GB4, Blenchmark, SPECint,…). The CB15 scores leaked now agree with the leaks contained in the MaxSun letter, where a 8-core Zen performed like a 6-core Broadwell
Tho Blender is one where Intel has had the upper hand forever. And they even refer it as base to tests their cpus for reviews. Even handbrake has been intel dominant. Bulldozer being slower in both then 2600k easy. In Blender losing to core i7 920 even. So it is clearly way better at least in these setups now.
You didn’t get my point. AMD used Handbrake in the marketing material of Bulldozer because it is one of the benchmarks where Bulldozer worked better, which hgave the false appearance that Bulldozer was better than it is really. And reviews showed that Bulldozer performed better in Blender than in other workloads.
AMD is using the same marketing trick now. They gave Blender and Handbrake numbers because in those special workloads Zen is closer to Intel, than it is really.
Yes, would be my guess. They are legit… and old ES samples. Screen shot shows a 3.1 GHz proc so the numbers are about what you would expect for an earlier version of the pre Ryzen CPU’s.
Even if they don’t quite match 6900K, price / performance will be interesting when pricing is officially announced
Stop spreading rumors, this is all fake. No CPU-Z, no actual solid proof. Put some Chinese letters with 16 Threads in english and let the fake stuff go!
We cant see the info on the system.
Everyone was speculating about amd new cpu, and they all failed cause the cpu is called ryzen. More speculation about the new vega saying its dual gpu. All i see is a bunch of Intel fanboys/employees spreading fake rumors. And the fact that tells me that amd is actually leaving the hole is their stock market evolution. Amd reduced their debt in almost 20% on the past 6 month, more money coming in, means more money for shareholders and for R&D of future projects, Amd already started working on the Ryzen follow up.
All i see is a benchmark of a stock 6900K vs overclocked 6900K. and there is nothing on the images that tell me otherwise. The images can be interpreted in many different ways. Ryzen was a very well kept secret, if something leaks it will be AMD to do it.
Those chinese leaked benchmarks are debunked already. it was revealed that they used an E5 2660 Xeon.