Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

youtube.com

Post reply on HN
Print view View original

31–40 of 78 posts

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#31

Post by neolefty »

vinw wrote: post #8
TwoNineA wrote: post #4

Preorders for the 9900K are 578$ USD. For that you can get a Ryzen 2700X, a nice motherboard and a 256GB SSD. Performance delta shouldn't be more than 15% deficit for the Ryzen for a few specific games.

Keep seeing this 'delta' word. From context it is basically like 'difference'. But is there a delta between difference and delta?

It comes from mathematical writing — the greek letter delta ∆ is often a variable used for the difference between two values.

So maybe you could translate it as "quantified difference".

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#32

Post by Symmetry »

vinw wrote: post #8
TwoNineA wrote: post #4

Preorders for the 9900K are 578$ USD. For that you can get a Ryzen 2700X, a nice motherboard and a 256GB SSD. Performance delta shouldn't be more than 15% deficit for the Ryzen for a few specific games.

Keep seeing this 'delta' word. From context it is basically like 'difference'. But is there a delta between difference and delta?

'Delta' has connotations of being quantitative that 'difference' doesn't have.

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#33

Post by gameswithgo »

nalllar wrote: post #24
gameswithgo wrote: post #13

Earlier quoted context omitted.

If you have a program that uses AVX instructions much the delta will hit ~2x

AVX512 on downclocks, you won't see the full 2x scaling. Cloudflare ran into some trouble with this: https://blog.cloudflare.com/on-the-dangers-of-intels-frequen...

AVX512 would be ~4x, but this intel CPU doesn't have it.

AVX2 is ~2x, Ryzen/AMD fakes AVX2 instructions with multiple SSE instructions.

Some AVX2 instructions downclock but not much, I see very close to 2x speedup over SSE2 with some workloads. Some of the downclock loss is made up for because there are more instructions available (gather, etc)

AVX512 might hit more than 4x improvement over SSE on some workloads despite the downclocks, due to all of the masking features. I have seen results consistent with this, 2nd hand. (I don't own an AVX512 cpu)

Anyway all of these things depend on workload, cpu, compiler etc. But it does happen!

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#35

Post by holtalanm »

asr wrote: post #12
cbg0 wrote: post #2

In text form: https://www.techspot.com/article/1722-misleading-core-i9-990...

This article says that Techspot has obtained the 9900 (without being bound by an NDA, as others are), and that Intel is releasing misleading results while other reviewers are bound by an NDA, but that Techspot is not going to show their own benchmarks--which could refute the misleading results Intel has authorized for public release--out of a sense of "professionalism." Actually , any "professional" journalism outfit…

the video doesn't dispute the benchmarks of the 9900, but rather they dispute the published compared benchmarks of the ryzen 7 2700X, which used default, unconfigured memory settings to push the benchmark down, while the 9900 benchmarks used configured memory settings (as it should)

Did you even watch the video?

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#36

Post by dragontamer »

nalllar wrote: post #24
gameswithgo wrote: post #13

Earlier quoted context omitted.

If you have a program that uses AVX instructions much the delta will hit ~2x

AVX512 on downclocks, you won't see the full 2x scaling. Cloudflare ran into some trouble with this: https://blog.cloudflare.com/on-the-dangers-of-intels-frequen...

AVX512 doesn't exist on the i9-9900k.

You need the "X-series" to get AVX512.

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#38

Post by dragontamer »

gameswithgo wrote: post #33
nalllar wrote: post #24

Earlier quoted context omitted.

AVX512 on downclocks, you won't see the full 2x scaling. Cloudflare ran into some trouble with this: https://blog.cloudflare.com/on-the-dangers-of-intels-frequen...

AVX512 would be ~4x, but this intel CPU doesn't have it. AVX2 is ~2x, Ryzen/AMD fakes AVX2 instructions with multiple SSE instructions. Some AVX2 instructions downclock but not much, I see very close to 2x speedup over SSE2 with some workloads. Some of the downclock loss is made up for because there are more instructions available (gather, etc) AVX512 might hit more than 4x improvement over SSE on some workloads desp…

> AVX512 would be 4x AVX2 is 2x, Ryzen fakes AVX2 instructions with multiple SSE instructions.

Not quite. Everything is fake, because everything is encoded into micro-ops first.

Ryzen's internal micro-op engine is 128-bit wide. But it has 4x pipes... each handling 128-bits at a time. So any 256 bit instruction will simply use two-pipes at a time.

-------

So the 256-bit instruction does in fact, execute at once.

The difference is that Intel has 3-pipelines, each of which can do a 256-bit instruction by itself.

-------

In effect: Ryzen is 4x128-bit pipelines, with the ability to merge pipelines together to do a 256-bit instruction.

Intel is 3 x 256-bit pipelines, with the ability (on Skylake-X) to merge pipelines together to do a 512-bit instruction

In any case, Intel has wider pipes than Ryzen. Intel Skylake is effectively a 256-bit CPU, while Ryzen is only a 128-bit CPU.

Re: Intel's New Low: Commissioning Misleading Core I9-9900K Benchmarks [video]

#39

Post by craftyguy »

m-p-3 wrote: post #27
asr wrote: post #12

Earlier quoted context omitted.

This article says that Techspot has obtained the 9900 (without being bound by an NDA, as others are), and that Intel is releasing misleading results while other reviewers are bound by an NDA, but that Techspot is not going to show their own benchmarks--which could refute the misleading results Intel has authorized for public release--out of a sense of "professionalism." Actually , any "professional" journalism outfit…

The thing is you just need to piss off an industry contact once to be taken off of their list, essentially neutering themselves in the future and making them less competitive compared to other websites. It's a sad state of affair, but that's the reality of it.

They were already neutered, they weren't given a part to review. the part they allegedly have was probably gotten through unofficial means, since it is highly unlikely intel would just give out parts to media outlets ahead of a launch without an NDA attached to it.
Post reply on HN