Crosshair VIII 2501's for testing

[Shamino]

still based on Patch C, should resolve some dram issues seen

https://www.dropbox.com/s/ojf8g8g1h7ot18j/ROG-CROSSHAIR-VIII-IMPACT-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/a4w0dghqe0qxvgy/ROG-CROSSHAIR-VIII-HERO-WIFI-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/1c3p11iloj3orbj/ROG-CROSSHAIR-VIII-HERO-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/o31pqe91pe15zm6/ROG-CROSSHAIR-VIII-FORMULA-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/24b0vppywj5r91b/ROG-CROSSHAIR-VIII-DARK-HERO-ASUS-2701.rar?dl=0

Edit updated links to 2701
bios 2701, names in links
fixes some bugs and raises settable fmax offset.
this is still based on the same AGESA version as 250x so dont expect any difference in Fclockiness against that version.

[78Staff]

78Staff wrote:
Just a follow-up on 2502... been running a bit less than a day, about 16-18 hours - similar results to 2501, in that I'm stable with 4 dimms and "standard" settings. Have done a bit of benchmarking, a bit of encoding, a bit of gameplay, lots of file transfers, etc. - so far so good.. and fingers crossed :).

I am not expressly overclocking - I am just running stock Optimized Default settings and enabling DCOP and setting FCLK to 1800 instead of auto. All other recent "workaround" changes trying to get stable with 4 dimms were not done, ie I have not disabled PBO Max, or DF Cstates, I have not under/overvolted anywhere, mem at 1.35v. Only changes other than mentioned are housekeeping related, ie enabling SVM for Virtual Machines, disabling onboard wifi and disabling Crate install.

Might not be much help or use to those that are trying to OC, but for my mainly content creation workstation so far 2501/02 seems to allow 4 dimm use with stability, which I wasn't able to do before since adding the 2nd set. Specs are in the dropdown, but basically C8H WiFi, 3900X, 4x16 GSkill 3600 C16 B-Die.

I'll test a bit more, I haven't let it sleep yet, so need to make sure no e41's when sleeping/coming out of sleep, and then might try to tune up the memory with the dram calc - but that has been a total fail for me so far :(. Maybe I should just keep DCOP timings and be happy I got there... Hoping this holds up as I plan to move to 5900 or 5950x as soon as I can get my hands on one.

Well got my first Event 41 restart on 2502 :(. Man I thought I had this licked since it's been up for over 24 hours, but I guess not yet.

So, I'm back to trying some of the suggestions from the OC.net thread - mainly disabling DF CStates, and manually setting memory speed to 3600, upping mem voltage to 1.37 vs using DCOP. Still keeping PBO Max enabled for now. Will report back...

[Silent_Scone]

78Staff wrote:
Well got my first Event 41 restart on 2502 :(. Man I thought I had this licked since it's been up for over 24 hours, but I guess not yet.

So, I'm back to trying some of the suggestions from the OC.net thread - mainly disabling DF CStates, and manually setting memory speed to 3600, upping mem voltage to 1.37 vs using DCOP. Still keeping PBO Max enabled for now. Will report back...

Is the memory stable? How are you testing it?

[the_patchelor]

any idea why 2501/02 has the FLCK "wall" / limitation again at 1900Mhz (like 2311)?
with 2402 (same AGESA Patch C) it was at 2066 with my CPU.

at least 2000 would be nice 😉

[xero404]

the_patchelor wrote:
any idea why 2501/02 has the FLCK "wall" / limitation again at 1900Mhz (like 2311)?
with 2402 (same AGESA Patch C) it was at 2066 with my CPU.

at least 2000 would be nice 😉

had the same issue for 2502 and could not set flck 2000 for my ddr4000 2x8 sticks. reverted to the 2402 beta and works again

[78Staff]

Silent Scone@ROG wrote:
Is the memory stable? How are you testing it?

Tested good with MemTest86 and MemTest64. I see several references to Karhu memory test but have not used it. I've also used Aida64 cache and mem benchmark although I'm not sure how valid is is as a testing tool.

IS this strictly a 4 dimm issue? My next test case will be to pull 2 dimms - I don't recall having any of these issues before going to 4 dimms, but was on older bios version as well... 2103 I believe. Worst case I can change to 2x32's, but I don't want to do that/spend the $$ if I'll face the same issues I am facing now with 4x16's. I also want to stay 59xx compatible as I plan on upgrading cpu.

EDIT - i went ahead and bought Karhu, seems to be recommended over the MemTest86. Will run it and report back.

And now that I think about it, before the Trident Neo I ran 4x8 of Ripjaws 3200 CL14 with no issues. I only upgraded to the Neo to get 3600/1800 FCLK, but I never had any issues with the RipJaws either... :rolleyes:

[the_patchelor]

why not simple staying with an older Bios version as long as you have the "old" Ryzen 3000 CPU?

the latest updates are mainly for Ryzen 5000 users to get the systems at least stable for now.

"never touch a running system" 😉

[78Staff]

the_patchelor wrote:
why not simple staying with an older Bios version as long as you have the "old" Ryzen 3000 CPU?

the latest updates are mainly for Ryzen 5000 users to get the systems at least stable for now.

"never touch a running system" 😉

valid point, and I am considering it as well, i even mentioned it earlier in this thread or maybe the other 4-dimm thread... 1302 seems to be the one that is guaranteed to work with 4 dimms without jumping through hoops, but every time I tried to revert to it I get a "not a valid bios file" error - I've re-downloaded it multiple times from the Asus site and get the same error.

Or do you have to use bios flashback to go back to that old of a revision?

[the_patchelor]

by default for older bios use USB Flashback

[78Staff]

the_patchelor wrote:
by default for older bios use USB Flashback

Thanks, I suspected that was the case. I've currently gone back to 2206 and am testing, I recall that being stable for me previously. So far I am able to post with DCOP settings which I could not do with the past few bios', and was able to run the 5-way optimizer for a gentle OC. Letting it run some encodes now and will see how she does.

[78Staff]

78Staff wrote:
Thanks, I suspected that was the case. I've currently gone back to 2206 and am testing, I recall that being stable for me previously. So far I am able to post with DCOP settings which I could not do with the past few bios', and was able to run the 5-way optimizer for a gentle OC. Letting it run some encodes now and will see how she does.

Two more E41's, both coming out of sleep mode :(. I think I am done with these Trident 4x16's for now. I'm going back to my rock solid RipJaws 3200 C14 kit and will research some 2x32 kits while waiting on a 59x0 to come my way...