Overclock.net banner
1 - 20 of 19441 Posts

· Registered
Joined
·
60 Posts
Anyone with 960 EVO (500GB) have such very low NVME benchmark scores? I am using the Samsung latest NVME drivers (and AMD chipset drivers) and high performance power plan on latest F10m for X570 Elite (with 3900x). Made sure drive is in PCIE 3.0 4x mode. Also tried with Windows native drivers, same numbers.

Any pointers? Same drive was running MUCH better on Z390/9900K system.
 

Attachments

· Registered
Joined
·
60 Posts
Got the same drive on an X570 Pro, benchmark while in use (MHW).

Edit: The first Magician and CrystalDiskMark screenshots were taken with the last MS Surface NVMe driver while the drive was in use. The second Magician screenshot was taken with the Samsung NVMe driver active, in idle.
Thanks for sharing. Your scores are very similar to what I used to get on my previous Intel based system. Any pointers on what I could be doing wrong? (I am using the slot above the GPU). Are you using any beta Samsung NVME drivers or the one posted on their website?
 

· Registered
Joined
·
60 Posts
Just a thought but did you do a clean install of win 10 or did you just take your old OS out of your old build?

Also maybe go in bios and force pcie to gen 3 instead of auto. There have been bugs if left on auto it will actually drop to gen 2 or lower which will cut your performance
Yeah did a clean install of course of version 1903. Also as per Crystal DiskInfo it is in PCI-E 3.0 4x mode. I will check all BIOS options once more to make sure nothing is obviously wrong.
 

· Registered
Joined
·
60 Posts

· Registered
Joined
·
60 Posts
I forgot to ask: 2x8 or 2x16?. I'm aiming for 2x16,but will they reach the target oc of 3600? Also, what timings? (I didn't download the calculator yet xD )

Enviado desde mi Mi A1 mediante Tapatalk
YMMV, but running 2x16 here no problem, I am too lazy to reboot to give you the timing, but they are exactly like in DRAM calculator (but I remember I ignored their 1.43v and kept 1.4v)
 

· Registered
Joined
·
60 Posts
With regards to the issue that users here are having with PCs not booting/BIOS resets, I replaced my OCZ ZX 1250w with a new Seasonic thinking that could be the problem and it still remains (the ZX 1250w had horrible coil whine so I don't mind the change one bit..)

My X570 Elite board now does one of three things.. (with my PSU left switched on 24/7)

* You press the power button, and get a Black screen.. the fans inside the case spin in a *certain way* the PC will go no further.. you can sit here for 5 mins and not boot or get a reset, so you power it off.
* You press the power button again and the monitor comes on and we boot up!

The 3rd scenario..

* You press the power button, and you get a black screen BUT you hear the fans inside the case ramp up, ramp down *quiet reset sound* , ramp up, ramp down *quiet reset sound* it does this once again then the screen comes on and the BIOS tells you that it has been reset.

Can we please have some acknowledgement from Gigabyte that this is a known issue that's being investigated for a fix?
This to me sounds like unstable RAM/CPU OC, if you have done any. I experienced the same when I was trying to max my RAM OC on many BIOS versions, rock stable once I figured it out and reached a stable RAM timing and voltage.
 

· Registered
Joined
·
60 Posts
recently finished my build: x570 Aorus Master, 3900x, 5700 XT and this memory kit: https://www.newegg.com/team-16gb-288-pin-ddr4-sdram/p/N82E16820331244

sadly the kit is not on the board's QVL (i remember checking a different x570 aorus board that i was planning to get for the build that did in fact list the kit as supported so i assumed the much higher tiered board would also support it). needless to say i've been experiencing some issues and i wanted to make sure these are caused by lack of compatibility and nothing else

basically, the ram will only run at what the system defaults to (minimum timings downclocked to 2400mhz, fabric at 1200) and at w/e ryzen master automatically sets the timings to when raising fabric clock in coupled mode.

the system does not boot in XMP (although the correct clock and timings are displayed when choosing the option in the bios), tried manually setting the dram voltage and then manually setting all the timings to no avail. i also tried manually overclocking in multiple clocks with very loose timings (based off of dram calc, also tried various dram calc timings at different frequencies) - but the system doesn't post. only thing that works is resetting cmos at that point.

on top of all of that, i've experienced similar issues (system not booting, more of less the same error codes) when setting the fan curve to silent.. had to reset cmos because of that aswell.

anyone have any similar experiences? is it possible this is all caused by lack of compatibility with the memory or is there a different issue?

any help would be appreciated. ty!
Congrats on your new build! Did you update the bios to the latest one? and did you install the RAM in the correct slots? (A2 and B2) This is what it says in your board manual to insure maximum compatibility for 2-dimms configs.
 

· Registered
Joined
·
60 Posts
Here's my first feedback about F20a... TLDR; I'm probably going back to F12a soon

The good:
  • Unlocked settings in AMD CBS
  • PMU pattern bits can be now configured manually
  • Better memory compatibility (can run CL16 with GDM Disabled now)
  • ProcODT at 60 with XMP profile fixed
  • Can now run VDDP/VDDG at 1000/1050
  • Better UI graphics...

The bad:
  • IMC is broken; memory bandwidth (Sandra) at 3800 MHz is much lower than F12a, almost same as when configured at IF 1800/3600 MHz.
  • Memory latency is different (Sandra); about 75% worse (2-5 ns), rest better (2ns)
  • Almost all voltages, LLC settings behave differently
  • Worse scores in benchmark, from a tiny bit to quite considerably, probably related to memory issues
  • Weird stalls under the main Windows install, feels much slower
  • Fixed SOC Voltage now is a problem, up to reboots during CB20 MT, have to keep in Auto
  • VDDP/VDDG at 900/950 is messy now, much lower scores
You nailed it!! was going crazy over CB20 reboots until I set the SOC voltage to Auto!
 

· Registered
Joined
·
60 Posts
Anyone found a way to set SOC higher than 1.080v without the system rebooting during CB20 run? I am facing that problem since F20a on the Aorus Elite. My RAM OC is not stable unless SOC is set to 1.12v which was fine in all BIOS versions prior to F20a!
 

· Registered
Joined
·
60 Posts
Try to change your "Over voltage" setting for SoC. I encountered a similar issue with reboots and I think that fixed it by upping it from AUTO, seems overly low in F20a @ stock.
I did try some other things as well but I think this one was what fixed the issue.
hmm, cannot seem to find that option in my Elite BIOS, do you mean a positive SOC voltage offset? I tried that but it has the same effect as setting it manually to the value I want.

Would really appreciate if you can expand on your answer a bit :) Thanks!
 

· Registered
Joined
·
60 Posts
Just want to mention "Over voltage protection" is set really low for F20a BIOS on the X570 AORUS XTREME.

AUTO = 300mv = ~1.100V SoC limit, you exceed this = reboot.
350mv = ~1.150v SoC limit, you exceed this = reboot.

So meaning, maximum 400mv over voltage protection only allows 1.200V maximum for SoC. They will have to fix this.

Unsure if Vcore is effected yet.
Indeed it is the case. Any way to be able to set that 400mv limit on the x570 Elite? :( Otherwise I think I will go back to F12.
 
1 - 20 of 19441 Posts
Top