does terrible in Steel Nomad as compared to any other non-RT test in 3DMark suite, or basically any modern game (for example, in PS5 ports my OC'd 6900xt is probably 20% faster than 3080ti/3090 and is (substantially) faster in many UE5 titles, and it's like 20% under in Steel Nomad). The only...
Another thing - the scores that i've published yesterday were done with CPU correctly boosting up to 5.75ghz, while GB shows (.gb6) that it was running at 5.5 ghz (which is actually the light multithreaded clock limit since some early-time AGESA). I checked it with hwinfo and it is congruent...
Rather, GB test loads are historically so short that CPU does not reliably boost to the max as the load stops before some cutoff is reached. It's not that in 'normal' (tm) tasks you have 10 ms long cpu loads and so on...
This is a very random "benchmark", here's two more results, I've run them one to the next with no reboots or settings changed:
https://browser.geekbench.com/v6/cpu/6668885
https://browser.geekbench.com/v6/cpu/6668824
Here's somewhat tuned 7950x3d results (my ccd1 is trash, can't clock above 5750 at all, surely dom can provide even better results on his binned CPUs):
Prefer cache:
https://browser.geekbench.com/v6/cpu/6665986
Prefer frequency:
https://browser.geekbench.com/v6/cpu/6668572
Are you running air cooler or just AIO? I was under impression that properly clocked 7950x can do about 1.4 gflops if thermals can be sustained (one of my friends tested 7900x with custom water loop at release time and he managed to score 1.1 gflops or about that)
What is AVX512? Is that thing that was supported by scrapped Knight Mill/Landing/etc, or Skylake-EP (with nice product segmentation with random silver xeons having one or two ports for "AVX512"), or mythical Cooper Lake and so on and so forth...
Shame Igor's lab name is already occupied :laughing: Well, we can always branch into Moor's Law is Alive, a trustworthy, good twin of Toms MLID youtube channel
I'm actually wondering - all these tests (as well as linpack if it is modded via usual 'reverse intel check' thing) actually use AVX-512, so this 40% uplift can be related to full-tempo AVX-512 :)
Is this JEDEC/XMP/Tuned ? My score is for either 6400 CL30-37-37-36 DR a-die or 7800 34-45-45-44 SR a-die, don't remember exactly which one (guess 7800, the date is fairly close to the time i've bought those sticks).
Above gflops score error can be fixed by just removing the checkbox in linx gui, then it'd run the next cycles (it's just a GUI limitaton, not real error)
According to my personal tests, x3d at relative minus 200 mhz is roughly equal to non-extra-cache cores in 100% membw sparse matrix multi algo, so it's not that much to be honest.
You don't really need the game bar, just set the driver (or bios) to 'prefer cache', it basically emulates one-ccd...
Yeah, it sounds in the same vein as "There is not enough space in the BIOS for Zen3 support on 3xx/4xx boards, here buy a new x570 board", which turned into a PR disaster for them later.
Thanks to hard FIT caps, those type of workloads don't generate much heat as maximum allowed voltage is just too low (below 0.9v ish). The highest power draw I saw was in cinebench 23 or 20, ~180 or 190W or so.
N7 process improvements, i guess, late Zen2 cpus (which were later separated into the XT SKUs) clocked much higher than the earlier Zen2 samples (my CPU that was bought at release date could barely sustain 4.5ghz/4.3 ghz OC with custom loop, while XTs could easily do 4.5 ghz all-core at much...
I suggest rereading what I've wrote - unless IO die or something that controls its registers has changed, even if your board can support 10000 mt/s, you'd still be stuck at 8000 mt/s or its whereabouts because the CPU cannot clock memory above 8000 mt/s without FSB overclock (which causes...
8000 mt/s for both 16gbit a-die and 24 gbit m-die. The thing is that the current IO die (or its firmware / w.e.) does not support mem multipliers above 80x, so even if your board and CPU can in theory do higher speeds, you'd not be able to do that unless you do bclk OC (which stops at...
No? Who told you that? If IO die hasn't changed, you won't see anything better than 8000 mt/s memory and 2200 ish IF (and i'm already using 7800 / 2166 on a 2dpc board)
It seems as if AMD tech PR has a KPI that forces them to do their best to lower initial sales and dampen hype. Not that I'd bought vanilla Z5 anyway, since it seems there won't be any bw improvements at all (so no extra perf for me).
It does not change (the clock), also it's the clock of the first core (and maybe 2nd or 3rd), CPU-z ignores CPPC and prevents you from changing that (unless you use special tools like BenchMate etc and launch it with predefined affinity)
Why you want em, if you are talking about normal TS, its CPU test is artificially capped at 16 threads and it's long a test of RAM performance rather than CPU
Here's a third theory for ya - the thermal density is so high, and AMD hasn't bothered with sanding the dies again (in addition to placing and gluing "structural silicon" on top of all dies), so a chiplet can't handle even 65W now at reasonable voltages and clocks.
N31 is extremely PL limited so you get high clocks in not that demanding games (including mostly CPU-limited MMOs) and in the other games and stress tests it drops to rock bottom or starts to eat upwards of 800W (check evc2se TS/TSE OC results at OCN). Either AMD aimed at 600W tdp initially or...
It's actually interesting that the 1T IPC numbers in one AMD barchart are lower than the nT performance uplift in the comparison with 14900k(s?), seems like either the frequency does not tank that much on Zen5 with multithreaded load as compared to Zen4 (5.1-5.2ghz on my 7950x3d with light sse...
Could be that the mobile versions are pushed back, who knows. Although some $3000 price therapy might force some people to rethink their credo of 'radeon exists to lower prices on geforce'
And what advantages does it give? It'd be a scorching volcano for dual-sided sticks, there's no real benefit in mem speed as compared to proper 1dpc boards (if you look at these slides with 2dpc empty slot nubs), overall RAM size etc is the same as the current SS/DS sticks
Same happened for Zen4 (and probably Zen3 too). I remember AMD released one number 13% something-something that caused the same crowd to implode (after hyping the release to death (in case of zen4)). Don't get the gloom, it's basically the same process, area size hasn't changed much, also there...
I think even low-end Asus 2DPC boards can easily do 7600 ish MT/s with one stick per channel, other brands (gbt/msi/asrock) can go even higher in general practice. It's all moot if IF is the same, which is the most probable outcome, I guess.
First of all, I really doubt it can do 2400 fclk. Also, in reality, you don't need expensive boards and RAM, just get something that works with HV less half-donkey-ed like my Gigabyte board does, and get one of Kingbank $100-150 hynix 24gbit/16gbit a-die kits that can easily do 8000+ with good...
What do you mean by "only"? If you compare results from the various outlets who tested either with JEDEC or AMD recommended specs (expo/xmp with more decent CL30 kits), it is actually disfavouring zen4 as compared to raptor lake(-r), especially in power-hungry nT tests (because enabling xmp/expo...
If the extra cache is again manufactured with dense process and glued the same way (looks like AMD decided to shave an extra 0.01c and just glue the structural silicon (tm) (C) on top of the cores ), highly unlikely in my opinion. Enterprise first, we peasants the last, duh.
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.