(Gigabyte X570 AORUS Owners Thread) - Page 213 - Overclock.net - An Overclocking Community

Forum Jump: 

(Gigabyte X570 AORUS Owners Thread)

Reply
 
Thread Tools
post #2121 of 2314 (permalink) Old 09-10-2019, 05:52 PM
New to Overclock.net
 
Join Date: Aug 2019
Posts: 15
Rep: 0
Quote: Originally Posted by JWMc View Post
I'd also like to report that the CPPC feature doesn't appear to work, that is if I actually understand what it purports to be. Is it not a feature designed to ensure that heavily loaded threads get scheduled onto the faster CCD / faster cores first?

I am using optimized defaults with CPPC set to Enabled, AMD Chipset drivers v1.8.19.0915, Windows version 10.0.18362.329, BIOS F7A / AGESA 1.0.0.3ABBA.

In the described configuration (and in all others, including BIOS F5 and F6), when I fire up CB R15 single the bench threads reliably land on my slower CCD, causing me to force them onto my faster CCD by setting the affinity mask. Otherwise I wouldn't have been able to report that I'm now sustaining 4.6ghz under single-core load using default settings with the new BIOS/AGESA. A lot of folks aren't going to be knowledgeable enough to figure out how to work around this issue, and of course we shouldn't even have to be burdened with it in the first place.
I haven’t seen this mentioned much, but while Windows does seem to prefer scheduling on my CCD1, it ping pongs between CPU1 and 2 for single threaded workloads. CPU1 is the fastest on the CCD, but CPU2 is the slowest.
This always seemed wrong to me, my assumption is it’s the Windows scheduler not respecting the actual core rankings in the CCD, but rather just preferring to keep threads on the same CCD/
I can see a marginal difference in Cinebench runs on my machine based on how long it spends running on CPU2 vs CPU1.
polygonhell is offline  
Sponsored Links
Advertisement
 
post #2122 of 2314 (permalink) Old 09-10-2019, 06:08 PM
New to Overclock.net
 
Join Date: Sep 2019
Posts: 2
Rep: 0
After many days of troubleshooting the problem of no post on my Aorus X570 Elite Wifi, the problem was incompatibility with the Intel Wireless Gigabit NIC card that comes with the Wireless Kit for the HTC Vive. On my Intel PC the NIC is working fine. Someone please notify this issue to Gigabyte.


https://www.intel.com/content/www/us...-overview.html


TY

Last edited by zeen; 09-10-2019 at 06:34 PM.
zeen is offline  
post #2123 of 2314 (permalink) Old 09-10-2019, 07:38 PM
New to Overclock.net
 
JWMc's Avatar
 
Join Date: Apr 2012
Posts: 15
Rep: 2 (Unique: 2)
Quote: Originally Posted by polygonhell View Post
I haven’t seen this mentioned much, but while Windows does seem to prefer scheduling on my CCD1, it ping pongs between CPU1 and 2 for single threaded workloads. CPU1 is the fastest on the CCD, but CPU2 is the slowest.
This always seemed wrong to me, my assumption is it’s the Windows scheduler not respecting the actual core rankings in the CCD, but rather just preferring to keep threads on the same CCD/
I can see a marginal difference in Cinebench runs on my machine based on how long it spends running on CPU2 vs CPU1.
My Cinebench R15 single score is roughly 15 points less when run on my slower CCD (~4.3ghz) vs my faster CCD (4.6-4.625ghz peak boost recorded on all 6 of its cores on BIOS F3), and about 20 points less for CB R20 single. My CineBench R15/R20 single-core score doesn't change for me *at all* beyond my 3600mhz CL16 XMP baseline as I go on up from 3600mhz CL14 through 3800mhz CL14. But remember, I'm artificially controlling my process affinity mask to force the single-core bench runs onto my fastest core. Outside of doing that it's very inconsistent as the threads end up bouncing around onto different cores having differing boost clocks.

But I can't draw any conclusions from my CB R20 multi scores so far. I'm wondering how much changes in room temperature are affecting it, and whether the A/C vent my open case is up against is pumping out cold air at the time or not. Water cooling must be a wonderful tool to get consistent thermal results during benching.

Ryzen 9 3900x (4.6ghz sustained CB R15/R20 SC) + x570 Aorus Master (F7A) + F4-3600C16Q-32GTZR b-die @ 3800mhz 14-14-16-14-28-42-298 @ 1:1:1 @ 1.51v, PDM off, GDM off, 1T
Cherry-picked outliers: CB R20 7422/521, CB R15 3273/209; Melody96: SA Chalk, Box Jade + Navy; GK707: HP 67g; Brown/Blue TKLs; Soon: ModelFKeyboards.com F77 in Beige Cast Zinc
Core i7 980X + EX58-UD5 (F13) + Samsung LP 6x4g @ ??
Sony X950G 55" ~1000nit HDR 4k TV as primary monitor @ 2160p60 4:4:4 + HP ZR24w + LG L246WP
JWMc is offline  
Sponsored Links
Advertisement
 
post #2124 of 2314 (permalink) Old 09-10-2019, 10:23 PM
New to Overclock.net
 
Dreams-Visions's Avatar
 
Join Date: Oct 2013
Posts: 9
Rep: 0
Glad to see the boosts are in line with expectations now.

I'll wait until a non-beta release. I'm not doing anything that makes getting an extra 75MHz necessary to mess with my current stability. Hopefully the final revision of this bios will be out in a couple of weeks.
Dreams-Visions is offline  
post #2125 of 2314 (permalink) Old 09-10-2019, 10:51 PM
- Insanity Beckons -
 
Elrick's Avatar
 
Join Date: Apr 2013
Posts: 3,905
Rep: 172 (Unique: 88)
Quote: Originally Posted by pal View Post
Currently using this with my Master series;

X570 AORUS Master - F7a

Elrick is offline  
post #2126 of 2314 (permalink) Old 09-11-2019, 02:20 AM
New to Overclock.net
 
BS Zalman's Avatar
 
Join Date: Aug 2019
Posts: 15
Rep: 0
I have tried using Beta Bios F57a for Aourus Master. Based on my experience on my 3900x a few core managed to hit 4.6 in iddle. When running C15 ST i get a small bump performance, from 206 to 208. Before the highest ST number i can get is 206, usually varrying between 204 - 206. But the core never managed to hit 4.6 Ghz when running C15, the highest clock is 4.50 which previosly only 4.25.

The problem with new Beta Bios is Memory OC is somehow degrading, with profile setting that previously stable in F6, give me a hard reset when stres testing memory with TestMem5. Even when i tried to lowering the frequency to 3733 it still give me a hard reset after a few cycle TestMem5.

So i rolled back to F6 and try to confirm the setting, running TestMem5 again with 1usmus_v3 profile, 10 cycles and the result are same as before it's rock stable.

Hopefully the public release Bios won't have this issue as well.
BS Zalman is offline  
post #2127 of 2314 (permalink) Old 09-11-2019, 06:45 AM
New to Overclock.net
 
Waltc's Avatar
 
Join Date: Jul 2019
Posts: 19
Rep: 0
Quote: Originally Posted by icf View Post
Hi,


Every time I shutdown the computer, power off the PSU, then restart the computer I get the following in the event viewer:


Event 29, Kernel-Boot
Windows failed fast startup with error status 0xC00000D4.


Event 6008, EventLog
The previous system shutdown at 8:48:37 PM on ‎9/‎10/‎2019 was unexpected.


Event 41, Kernel-Power
The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.


3800x, x570 Aorus Master F6, 970 Evo plus 1TB M.2, Windows 10 1093, latest chipset drivers etc.

First things first--turn off *fastboot*--both inside Windows if you have it turned on, and in the bios if you have enabled it. If you boot from SSD or NVMe you just don't need it, not to mention that it's always been problematic for as long as I can recall. I post and boot to the Win10 desktop in ~10 seconds, cold boot or warm, makes no difference. Fastboot is a hangover from the HDD days, when it could take 90 seconds-120 seconds to get to the desktop.

Windows10x64 v1903, build 18362.10019, 19.9.2 Adrenalins, Aorus Master x570 UEFI bios F7a, Corsair HX-850 PSU (72a x1 12v rail), Ryzen 5 3600X @ 3.8GHz & boost, all cores @ def v, Stock AMD cooler, 2x8GB, Patriot Viper Elite PV416G320C6K @ XMP 3733MHz 18 20 20 42 1T, LG MultiDrive DVD writer SATA, Boot: Samsung 960 EVO NVMe 250GB, 2x 1TB WD Blue 7200 rpm, ST2000DM S3 2TB, ST4000DM004 S3 4TB, BenQ EW-3270U 4k, HDR10 monitor, AMD 50th Anniv. 5700XT 8GB GDDR6
Waltc is offline  
post #2128 of 2314 (permalink) Old 09-11-2019, 06:59 AM
New to Overclock.net
 
Waltc's Avatar
 
Join Date: Jul 2019
Posts: 19
Rep: 0
Quote: Originally Posted by GBT-MatthewH View Post
Hey guys, finally back in the office. The 1st iteration of 1003 ABBA is out in the wild, however (for now) I am going to hold back posting links until I can talk to our team in HQ.

Tweaktown seems to have a mirror up for those interested. Obviously main change log is improved boost speeds, but we also opened up spread spectrum for those who have been asking Its gonna be a few days for me to catch up on the thread, but ping me if you find anything urgent.

Matt, Assuming the beta bios is F7a...? Spread spectrum introduced a toggle with F6, I believe.

Windows10x64 v1903, build 18362.10019, 19.9.2 Adrenalins, Aorus Master x570 UEFI bios F7a, Corsair HX-850 PSU (72a x1 12v rail), Ryzen 5 3600X @ 3.8GHz & boost, all cores @ def v, Stock AMD cooler, 2x8GB, Patriot Viper Elite PV416G320C6K @ XMP 3733MHz 18 20 20 42 1T, LG MultiDrive DVD writer SATA, Boot: Samsung 960 EVO NVMe 250GB, 2x 1TB WD Blue 7200 rpm, ST2000DM S3 2TB, ST4000DM004 S3 4TB, BenQ EW-3270U 4k, HDR10 monitor, AMD 50th Anniv. 5700XT 8GB GDDR6
Waltc is offline  
post #2129 of 2314 (permalink) Old 09-11-2019, 07:46 AM
New to Overclock.net
 
Martin778's Avatar
 
Join Date: Feb 2012
Location: the Netherlands
Posts: 1,371
Rep: 15 (Unique: 15)
I don't understand, in games the clock is constantly bouncing between 4375 and 3600. All the time.

Core i9 9900K @ 5.2 || EVGA Z390 DARK || 2x8GB G.Skill TridentZ RGB 4600 C18 @ 4266 C17 || GPU: EVGA Kingpin RTX2080Ti @ 2130MHz || Samsung 970PRO 1TB + 860EVO 2TB || Lian Li O11 Dynamic White || Corsair AX1600i || Screen: LG B8 OLED 55 and LG 34GK950F
Corsair H150i PRO + Noctua NF-A12x25
Martin778 is offline  
post #2130 of 2314 (permalink) Old 09-11-2019, 08:56 AM
New to Overclock.net
 
Nopileus's Avatar
 
Join Date: Dec 2011
Location: Germany
Posts: 102
Rep: 4 (Unique: 4)
On my 3600 + x570 elite with the ABBA BIOS there appears to be a loss of Multicore performance in Cinebench R20, i've lost 100-130 points on my previous PBO result (3570-3600 compared to 3700 points), however the Singlecore clocks seem more consistent now with 498 on a dirty run (not all background applications off) compared to 490 on a clean run prior on the old AGESA.

This is both with PBO enabled but even my stock multicore results were around 3650 points so this is a bit mediocre.

Likely not something noticeable in use but interesting regardless.

Edit: Max clock with PBO hasn't changed, still spiking at 4350 briefly at low loads but more typically 4300mhz

Last edited by Nopileus; 09-11-2019 at 08:59 AM.
Nopileus is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 21 (3 members and 18 guests)
Cata79 , IamEzio , nangu
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off