In Cities Skylines 2 thread loading at worst is around 60% with 8c/16t CPU(5800X3D). This means unless you have less than 6 cores(w/SMT), it's all about single thread perf. Buying more cores for idling makes no sense whatsoever. And the game is really GPU heavy in current form.
The other CCD is running at zero MHz. Because cores on it are powered down(parked). 3rd party monitor apps can't show actual clock speed, they poll cores for information and it would wake up them. So they report last known clock speed which is typically base clock(before core going dark)...
RX 580 doesn't have proper hardware to decode VP9(or AV1). No drivers will fix that. You need minimum RDNA1 dGPU or Ryzen APU(VCN 1.0) for VP9 HW decoding. RDNA2(VCN 3.0) is minimum for AV1 HW decoding.
My RX 6800 decodes completely fine 4K VP9 and AV1. I tested even two 4K@60fps YT videos...
AMD GPUs were most energy efficient of last generation. Hands down. 5000-series wasn't bad either, about equal to Nvidia. AMD is about to launch 3rd generation after last GCN-based furnace, this thought of AMD being power hungry should die already.
Thermal cycles can squeeze paste out in certain conditions. Doesn't mean the paste is bad, just not good fit. I'd try something with higher viscosity, they tend to "stick" better. Though cooler removal after longer time needs to be done with care(warmed and with twists).
These numbers from the wall align very well with 142W power limit. 5800X at stock has 142W PPT, cores get around 120W(idle package power is higher than SoC power). VRM+PSU efficiency is about 80%. 120W/0.8 = 150W. 150W + 52W = 202W.
110W is just power consumption of CCDs, not whole CPU. "103% of 142W" is 146W. It's not 110W. And TDP doesn't really even exists, it's meaningless number chosen by AMD in relation to PPT. For example 5600/X PPT is just 17% higher than TDP. Threadripper and Epyc TDPs are same as PPT. Odd thing is...
Mfg date on the CPU is in year/week format. I checked a few reviews from launch day and almost all had week 38(mid Sep) 5950Xs, a couple week 35(late Aug). I bet your 5950X's are 2038's too, week 18 is too early.
Not exactly apples to apples but very close. 12900k and 5900X with DDR-3200 are around 150fps. Doesn't look good for expensive high end memory kits though, this type of tests are biggest selling point of high performance memory.
Average of three benchmark runs is 228fps, in blue text. You are looking at wrong numbers, those are real time numbers constantly updating. 290fps is simply framerate when screenshot was taken, not max. Obviously very lightweight spot.
The gap is what I said 3300X vs 3600/3700X/3900X/3950X and bunch of other skus in-between. Sorry, pointing out one barely existing crippled Zen2 with halved L3 doesn't prove me wrong. And most certainly not other CPUs with different architectures. I won't continue this further.
10% higher clock speed. L3 size is doubled, 8MB to 16MB. Take those away and how much there is actual difference left. Not much considering 2+2 is worst possible configuration. 3600 in same test pretty much matched 3300X. These have same 16MB L3 and similar clock speed.
Not denying any of this however it doesn't seem to be relevant when benchmarking games. Multiple CCXs doesn't really hurt Zen2 either. In best cases single CCX variant is couple percent faster than rest, it's meaningless. Even 5700G is barely faster than "lowly" 3700X despite having higher clock...
Pictures are nice but threads in reality don't behave like that. Because Zen L3 is victim cache, it only contains data pushed out from L2 cache. Data in L3 is from threads running in that particular CCD. It also means it's unlikely the L3 on one CCD contains anything needed on other CCD. Also...
It says "I'm cherry picking to make my argument". Also no one should buy 4GB card for AAA gaming if you are not willing to do big compromises. This was true 5 years ago.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.