• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

【Maxon】Cinebench R20 Benchmark Thread

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
.vodka said:
If so then I'll just have to encrypt it and share a password, and name the file something less obvious.

It's ridiculous they're using the store to distribute a classic windows app.
Click to expand...
Using windows to distribute a classic windows app?!?! Those crazy kids with their foolish ideas.

Also how is cinebench a classic windows app?
 
TheELF said:
Using windows to distribute a classic windows app?!?! Those crazy kids with their foolish ideas.

Also how is cinebench a classic windows app?
Click to expand...

It's not UWP based.

You take CB R20's folder right as it's installed in W10 and it runs on W7/8.1, no errors, no complaining, nothing. Hell, I wonder if it runs on Vista or even XP.

If you're going to distribute that through a store, then you might as well just provide a link to a standalone distribution as they'd done with every CB before.

Anyway, enough offtopic.
 
Last edited:
.vodka said:
It's not UWP based.

You take CB R20's folder right as it's installed in W10 and it runs on W7/8.1, no errors, no complaining, nothing. Hell, I wonder if it runs on Vista or even XP.

If you're going to distribute that through a store, then you might as well just provide a link to a standalone distribution as they'd done with every CB before.

Anyway, enough offtopic.
Click to expand...
Oh ok, because older windows aren't even able to run the store but the software does run on older windows ...the word classic confused me.
 
Even though Maxon claims that R20 is an "AVX" application, there are no 256-bit instructions in it. And that's great news for Zen and Zen2!

This is the most costly function (ray-triangle intersection). It contains some FMAs, which theoretically penalize Zen, but the ratio of FMA to load is too low for it to be the bottleneck (Zen can do 2 128-bit load + 2 128-bit FMA).
Code: 
vmovups xmm0, xmmword ptr [r11+rsi*1+0x20]
vmovups xmm2, xmmword ptr [rdx+rsi*1+0x20]
vmovups xmm3, xmmword ptr [rax+rsi*1+0x20]
vmovups xmm4, xmmword ptr [r15+rsi*1+0x20]
vmovups xmm5, xmmword ptr [r14+rsi*1+0x20]
vmovups xmm5, xmmword ptr [r13+rsi*1+0x20]
vfmsub213ps xmm0, xmm13, xmm9
vfmsub213ps xmm2, xmm12, xmm10
vfmsub213ps xmm3, xmm8, xmm11
vfmsub213ps xmm4, xmm13, xmm9
vfmsub213ps xmm5, xmm12, xmm10
vfmsub213ps xmm1, xmm8, xmm11
vpmaxsd xmm0, xmm0, xmm2
vpmaxsd xmm3, xmm3, xmm7
vpminsd xmm4, xmm4, xmm5
vpminsd xmm5, xmm1, xmm6
vpmaxsd xmm0, xmm0, xmm3
vpminsd xmm0, xmm0, xmm3
vpcmpgtd xmm0, xmm0, xmm3
vmovmskps edi, xmm0
xor rdi, 0xf
jz -112

Every other function is only scalar code, which still executes twice as fast on Zen as on low-IPC Intel cores.
 
i7 8700 non k 4.3Ghz all core turbo 2x8gb dual channel 2666mhz ram

3461 multi score

Been ages since i ran a synthetic benchmark.As more of a gamer,i guess these don't appeal much to me.Needs more cowbell?
 
amdfan111 said:
Even though Maxon claims that R20 is an "AVX" application, there are no 256-bit instructions in it. And that's great news for Zen and Zen2!

This is the most costly function (ray-triangle intersection). It contains some FMAs, which theoretically penalize Zen, but the ratio of FMA to load is too low for it to be the bottleneck (Zen can do 2 128-bit load + 2 128-bit FMA).
Code: 
vmovups xmm0, xmmword ptr [r11+rsi*1+0x20]
vmovups xmm2, xmmword ptr [rdx+rsi*1+0x20]
vmovups xmm3, xmmword ptr [rax+rsi*1+0x20]
vmovups xmm4, xmmword ptr [r15+rsi*1+0x20]
vmovups xmm5, xmmword ptr [r14+rsi*1+0x20]
vmovups xmm5, xmmword ptr [r13+rsi*1+0x20]
vfmsub213ps xmm0, xmm13, xmm9
vfmsub213ps xmm2, xmm12, xmm10
vfmsub213ps xmm3, xmm8, xmm11
vfmsub213ps xmm4, xmm13, xmm9
vfmsub213ps xmm5, xmm12, xmm10
vfmsub213ps xmm1, xmm8, xmm11
vpmaxsd xmm0, xmm0, xmm2
vpmaxsd xmm3, xmm3, xmm7
vpminsd xmm4, xmm4, xmm5
vpminsd xmm5, xmm1, xmm6
vpmaxsd xmm0, xmm0, xmm3
vpminsd xmm0, xmm0, xmm3
vpcmpgtd xmm0, xmm0, xmm3
vmovmskps edi, xmm0
xor rdi, 0xf
jz -112

Every other function is only scalar code, which still executes twice as fast on Zen as on low-IPC Intel cores.
Click to expand...
nice to know
the results are in line with that

just what Maxon did with the benchmark after 5 years ? 0,0x AVX ? I am dissapointed
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top