5600X > 5600G. Explains the lower cost to the main Zen3 CPU.
Also, as a comparison with the 11400, 5600G has a faster GPU but performs worse overall when using a discrete card. It doesn't make it a worthy buy for gaming systems where folks are holding on to integrated graphics while waiting for RTX prices to drop.
Also for X570, there's also a bug with WD's PCIE4 NVME drives having horrible speeds when running on M2 slots connected to the PCH. It's probably just WD though, no reports on the same issue happening on Phison controller SSDs or Samsung's 980 Pro.
Upgraded my CPU cooler to Corsair iCUE H150i Elite Capellix..
Temp about to settle down at 70'c under Prime95 stress test. However, the clocks speed for the 3700X only at 37xx mhz when im doing stress test. Is this consider normal?
I doubt VRMs are toasty given the mobo in your signature. Is the CPU only at stock clocks?
Technically your 3700X is within spec here, it's advertised all-core speed is 3.6GHz after all. Try enabling PBO2 or setting a manual all-core overclock.
How well do other thermal pastes fare against NT-H2 if I wanted to avoid using Kryonaut? E.g. Gelid GC Extreme, Arctic MX-5, etc. Planning to swap coolers with an AIO on my end.
as long the cpu cooler is rated for your cpu max boosted tdp, 3 degree plus minus doesn't affect much la…
I'm more interested to see how little I can avoid a loss of thermal efficiency when switching from Kryonaut to another thermal paste, while avoiding Kryonaut's issues with longevity. Repasting my RTX 3080 with Kryonaut gave me an increase in clockspeeds, and swapping my 5800X from NT-H1 to Kryonaut gave it around 2-3C more headroom.
NTH2 is better because it doesn't degrade as quickly as Kryonaut, also won't suffer from pump out issue (this is an issue for Zen CPUs because the IHS is convex). I have used Kryonaut since 2014 and my experience is that you need to change them out in about ~1 year or so. NTH2 has been more than 1 year in my 3900x system, still hasn't suffered any degradation. On my 5900x system, I'm still using the Kryonaut on it but these few weeks I have noticed higher peak temps. No, it's not because of weather affecting ambient temp either because my room temp is always controlled at 22c whenever PC is turned on.
That said, they both perform exactly the same temps wise when both are freshly applied. 0.5c difference is NOT a difference ya.
Ah, exactly what I wanted to know. Only compared NT-H1 to Kryonaut and did see a temp difference with that. If NT-H2 has no discernible difference I guess I can use that.
Tier 2: Intel gen8/9/10 without B560 or Z series chipset Ryzen 3000
Tier 3: Ryzen 4000/2000/1000 Intel gen7 and earlier
Ryzen 3000 used to be tier 1 budget options, only because of the lack of B560 chipsets and higher Intel CPU prices earlier making them less price competitive.
Actually don't have high hopes for their 3D v-cache either. Zen 3 is complicated enough as is, just tacking on massive amounts of L3 isn't going to alleviate the problems it already has. This is just a stop gap for AMD so they can have something to compete with ADL (eventhough it's a futile attempt) so that they have a buffer before launching Zen 4.
Probably enough to claw back the lead in some gaming benchmarks (some titles benefit a lot from memory latency and show huge benefits on Samsung B-die with low timings or specialty CPUs with high cache like Intel's Broadwell quadcores), no improvement in other tools. Might be harder to cool too.
It's made for people who are on Zen 2 who has a good X570 board, good b-die DDR4s and not wanting to wait until Zen 4 releases in Q4 2022 because FOMO.
Honestly, it's a pointless release just like Coffee Lake Refresh was.
QUOTE(Bonchi @ Nov 11 2021, 03:37 PM)
Or rocketlake
Coffee Lake Refresh has 2 extra cores, RKL added PCIE4 and AVX-512 for those niche users who wanted it.
Yeap if they did that then we would have raptorlake today and alderlake last year lol.. but owell, when you have a bad ceo, no matter how good your engineers are, it wont be fruitful.
If Intel wasn't too ambitious with their 10nm process with their failed experiments on cobalt etc:
Q4 2017 - Cannon Lake instead of Coffee Lake 2018 - Ice Lake desktop instead of Rocket Lake 2.5 years later 2019 - Tiger Lake 2020 - Tiger Lake shrunk to Intel 7nm 2021 - Alder Lake on Intel 7nm, not "Intel 7" that's actually their 10nm process
To what extent is the issue due to their old CEO and not their fab chiefs? I don't know.
Sounds like the extra 3D V-Cache on top of the cores is worsening heat transfer. Separating the I/O die from the CPU already negatively impacts heat transfer to the IHS, and is why Zen3 desktop CPUs run hotter than Intel's when at the same power draw.
Marketing now out. AMD claims 5800X3D is faster than 12900K in gaming. I won't be surprised - Anandtech did a test on the 5775C many years after Broadwell launched. Even with the abysmal max clockspeeds on that CPU vs Haswell, the CPU was faster in gaming than any *lake quad core chip. All thanks to that massive L4 cache.
But Alder Lake should handily hold its advantage over most productivity benchmarks anyway... so the 12700K etc would still be a better buy for most people.
This post has been edited by yimingwuzere: Mar 15 2022, 09:39 PM
Secondary PC with a 3700X died. Sent it for warranty claim, was told it was OOS. They gave me options to pick either a 5600X or to top up RM300 for a 5800X.
Should I top up or go down to 6 cores? This setup only has a Wraith Prism for cooling, so not sure if going to 8 is worthwhile without tweaking power usage with Hydra/CO etc.
Obviously 5800x is the better choice since it's only RM300 extra.
But the issue with the single chiplet Zen 3s is that it gets stupidly hot. Your wraith prism will not be enough to handle it without running at constant 100% fanspeed and with PBO disabled, which kinda defeats the purpose of the 5800x gaming prowess to begin with.
So if you're going for the 5800x, do pick up a better cooler to go with it.
Yeah makes a lot of sense for me. Guess I'll go 5800X then. I can probably just use Hydra and CO/PPT/EDC/TDC to tweak its power usage.
OTOH the 5700x just launched for RM1.3k, maybe you can get that for a 1:1 exchange? Would be a better choice than the 5800x IMO.
I wasn't given the option to pick a 5700X.
Given that the 5800X should be a better bin overall (Gamers Nexus reviewed the "5800" OEM CPU with the same specs as the 5700X and seemed to conclude that too) I think I'll try my luck with tuning it.