dual AMD EPYC Rome (7nm/64 Core)
running at 1.8ghz
discuss
AMD EPYC Rome (7nm/64 Core)
Jaxson Adams
Nathan Ortiz
Too bad cinebench doesn't run under Linux. If WIndows shits itself with 32 cores it probably does even worse with 64.
Nolan Foster
>128c/256t
wat
Evan Peterson
>Cinnebench
>low power ES
Yeah, that's a no from me
Matthew Moore
its windows 8 tho for some reason
Dylan Thompson
>What is Hyperthreading?
Oh right, intel lost even that.
My bad
Jordan Davis
Great, a processor dedicated to ... uh, waving your Cinebench dick and unpacking 7zip archives I guess.
Luke Martin
its 2x 64core romes
Matthew Thompson
>discuss
You're a fucking slowpoke
Asher Walker
are you saying that 64c/128t=128c/256t?
Robert Perry
That's like 2 months old eng sample
Jayden Butler
Why can't you read?
Gabriel White
...
Carson Flores
First of all, cinnebench is too short to test this many cores, a bunch of them don't even get to work, cinnebench should be at least 500% longer
Secondly, you posted old ES that might just be low power versions since we already have Rome 2.3GHz numbers.
Thirdly, turbo is not even working on these chips, even if it was the score would practically be unchanged since the benchmark doesn't last enough for the turbo to properly kick in.
Blake Nelson
for cpu rendering Rome is going to dominate. Hell even Threadripper is starting to change the considerations of GPU rendering.
Justin Gomez
In fact over some 30 cores on CB it's better to disable turbo if you want consistent scores, turbo can introduce up to 20% variance in score on high core counts, CB has no problem with this under 20 cores.
Tyler Lee
8.0 > 10 > 8.1
Camden Perry
what are you a corelet?
Tyler Cruz
threadlet detected.
if leak is true, you're looking at a 2 socket solution for ROME so 2x 64c/128t
Mason White
correct
and for comparsion intel cascade lake 48 core 96 thread got about the same score but was running on 2.50ghz instead of 1.8
Nathan Anderson
Too bad Windows 8 doesn't have as small Window borders as Windows 10 has.
Nathaniel Murphy
>benchmarking NUMA systems on Windows
Tyler Howard
>Genuine Inrel
>Inrel
>In R el
Hudson Ross
>Comparing a $20,000 CPU to a $2,000 CPU
Noah Roberts
H O L O C A U S T
Brody Martinez
OY VEY
Evan Cook
>2 Glue™Technology™ Skyl..eeh Cascade Lake server Chips
>20k
thats..for one glued part, right? Intel and the Investors love their high profit margin, i highly doubt that part is 'only' 20k
Adrian James
>running on 2.50ghz instead of 1.8
i'm not taking those reported clocks for bare coin and will wait to see until actual release presentation. but not too bad for an Engineering Sample from AMD
Jaxson Ross
>inrel
Fake shit lul
Jaxson Clark
>1 off