- Mar 3, 2017
- 1,687
- 6,243
- 136
I wish I could but I don't think I will get anymore screenshots.Can you arrange such a test in AIDA64? It is more important than FP32/64 and AES
TF-GrayWizard on techpowerup explained how to manipulate cpuz scores. For Ryzen run the avx512 beta test then change back to the version 17 test, but don't run it and you can choose whatever you want to compare with. It will keep the avx512 results. Heres a 7950x3d faster than a 7950x and 13900k. I don't have my 14900ks installed anymore so I can't see if it can be faked easily with the various avx tests.That doesn't look good for Zen 5: https://www.guru3d.com/story/intel-...thread-performance-improvement-over-i913900k/
performance hit on Windows comes from VBS being enabled, if you have it
Nah, it's the famous IO kernel bug that's been present since 2.6 released some 20 years ago, it was memed to death since then (just google "Bug 12309")Can you point me towards a guide how to replicate the Linux IO issues? Personally, I have never encountered such issue. It feels like an implementation issue of Files/Dolphin/Nautils.
Ask him to lay of the LN2I asked my guy about it, he said he was just messing around with it. He re-ran it without touching anything.
View attachment 102402
In my eyes this is clearly not a legit run..I am 100% certain the screenshot is unaltered
I am 100% certain the screenshot is unalteredAsk him to lay of the LN2
Does the distance look correct here ?
In the screenshot above the cpu string name seem to be centred..
View attachment 102404
My screenshot taken from my own video here: but i also ran a much older version of CPU-Z so maybe things have changed (?)
Can your guy do a Rapydmark (High setting) run? Please share the screenshot and the text results if he does.I am 100% certain the screenshot is unaltered
Not sure that should happen outside of placebo, the effect on geekbench scores is just couple percent IIRC. I never really looked into testing the impact though, I just have it on all the time (except when I wanted to turn it off for such benchmarks).very true, disabled it (core isolation/memory integrity) and everything feels way snappier
It's definitely a falsified benchmark. The screenshots are completely unaltered though. Entirely done by tricking/glitching CPU-ZOnce again, CPU-Z is an incredibly easy benchmark to fake.
It's not that useless. It's the quickest way to see the effect of overclocking. And its AVX2 test is kinda dangerous coz it caused a very audible and ominous coil whine on my server mobo. No other benchmark I've run has done that. Don't underestimate it.Use CPU-Z for what it was intended for, and ignore the useless "Bench" tab.
Well that SKU is probably a successor to the cheaper 6 cores Zen4 units, so if we look in that perspective there could be improvement.That's just a pro sku, you won't see it in any store. AMD released it early to cash in on the AI hype in the corporate sector
The regular Ryzen 71360 probably uses kraken die (4+4, 8CU), and will be released much later
It's a successor to big phoenix (8c), not small phoenix. They reduced the number of CUs to fit a 50 tops npu, and upgraded the cores from Zen 4 to a mix of Zen 5 and Zen 5cWell that SKU is probably a successor to the cheaper 6 cores Zen4 units, so if we look in that perspective there could be improvement.
Phoenix and Strix Point are both "Premium Mobile" parts according to AMD (Fire Range being a desktop replacement part and Strix Halo a completely new segment for the X86 market). Of course Strix will target also higher performance (and thus price) but there is some overlapping. The SKU in the test can be completely substituted by the top Kraken Point part, which will be identical in performance and power, even higher because the Ryzen 160 tested appears to be a 3+5 part instead of the 4+4 in Kraken (KP die is not out yet, of course). Graphics seems likely to be cut down on both compared to top Strix, this 160 PRo part may even have less CUs than top Kraken as someone says it will be a 12 CU part. It seems they will have the same NPU. So why the Kraken Point should be a successor the the 6-core Phoenix but the same/lower performing cut-down Strix cannot?It's a successor to big phoenix (8c), not small phoenix. They reduced the number of CUs to fit a 50 tops npu, and upgraded the cores from Zen 4 to a mix of Zen 5 and Zen 5c
Pretty sure Kraken is also supposed to be 3+5 and the bin 2+4 ( -1 , -1).3+5 part instead of the 4+4 in Kraken
Kraken is 8 CUs. Bin is 4CUs.Graphics seems likely to be cut down on both compared to top Strix, this 160 PRo part may even have less CUs than top Kraken as someone says it will be a 12 CU part.
Pretty sure all the rumors so far gave Kraken as a 4+4 and not 3+5.Pretty sure Kraken is also supposed to be 3+5 and the bin 2+4 ( -1 , -1).
Kraken is 8 CUs. Bin is 4CUs.
Yeah, 3+5 is bizarre.Pretty sure all the rumors so far gave Kraken as a 4+4 and not 3+5.
More bizarre than these?Yeah, 3+5 is bizarre.