Overclock.net - An Overclocking Community - View Single Post - (Gigabyte X570 AORUS Master Overclocking Thread)
View Single Post
post #50 of (permalink) Old 08-25-2019, 07:11 PM - Thread Starter
Heuchler
New to Overclock.net
 
Heuchler's Avatar
 
Join Date: Mar 2012
Posts: 2,031
Rep: 274 (Unique: 193)
@ l0rdraiden - Glad your vcore is being applied correctly. Voltages not being applied properly could be the most annoying issue. I believe that motherboard partners can't change the AMD CBS part of the BIOS. So some settings a redundant in the BIOS and if they don't match it probability isn't a good thing.
The first reddit post you link the OP doesn't see temps over 80c because how the Ryzen 3000 [Matisse] new boost algorithm handles things.
If your temps are too high it won't clock high. But he is assuming because he is at 80c that he has thermal headroom left.


Quote: Originally Posted by The Stilt View Post
Matisse Boosting Algorithm

Setting the thermal limits below stock (95°C) make no difference, since the boost algorithm already uses lower limits.

The original limits for Ryzen 3000 SKUs were:

- 3600 = 4100MHz (80-95°C) / 4200MHz (< 80°C)
- 3600X = 4200MHz (80-95°C) / 4400MHz (< 80°C)
- 3700X = 4200MHz (80-95°C) / 4400MHz (< 80°C)
- 3800X = 4300MHz (80-95°C) / 4550MHz (< 80°C)
- 3900X = 4400MHz (80-95°C) / 4650MHz (< 80°C)

Since then, it appears that the HighTemperature limit has been reduced further to 75°C (from 80°C).
New SMUs also have introduced "MiddleTemperature" limit, but that gets disabled when PBO is enabled.

HWInfo is also able to display these limits (fused values).
https://www.overclock.net/forum/28085580-post549.html
SMU firmware = System Management Unit (includes Power Management among many other things). I started reading up on it a little.


@Dreams-Visions - 2xCCD Matisse Chip (3900X and 3950X) seem to prefer 4 single rank DIMMs over 2 dual rank DIMMs

Quote: Originally Posted by The Stilt View Post

For 2 CCD SKUs, 2 DPC SR configuration seems to be the way to go.
Both the 3600 and 3700X did 1800MHz UCLK on 1 DPC DR config, but most likely due to the discrepancy of the two CCDs in 3900X, it barely does 1733MHz on those DIMMs.
Meanwhile with 2 DPC SR config there is no issue in reaching 1866MHz FCLK/UCLK. That's pretty unfortunate since 8GB DIMMs cannot be considered as desireable or future proof as 16GB ones, due to the 32GB limitation.
Sure 16Gb ICs exist nowdays (hence allowing 16GB SR modules), but none of them can come even close to B-die in terms of frequency and timings.



Increasing cLDO_VDDP seems beneficial > 3600MHz MEMCLKs, as increasing it seems to improve the margins and hence help with potential training issues. On previous gen. products it was only useful for shifting the MEMCLK holes, which were
present on certain CPU, motherboard and DIMM combinations. But then again, we never did this kind of MEMCLKs on those parts.
Damn good question. Found another X570 Master owner running Hynix CJR SR and he was testing 4x8 at 3800. Before I was thinking about picking up a 32 GB Micron Rev.E based kit. Samsung B-die still seem to be kind. Micron Rev.E
seems to have some advantages over Hynix CJR. On the other hand CJR seem to able to have lower tEFC and tFAW. Nighthog has his Rev.E memory at 4266. Seems like a lot of fun to overclock it.
I played around Flare X based B-die on Ryzen 1000-series and the motherboard wouldn't even do rated speed (not a fan of that motherboard company). 3440x1440 is about 20-25% performance hit over standard 2560x1440
from my testing depending on game and settings. 720p and 1080p FPS differences between the top three IC (B-Die, Rev.E and CJR) ComputerBase link a few post before. I wouldn't really worry about it too much
and spend the difference on getting a better GPU or put it aside for a future GPU upgrade. Huge L3 cache seems to be doing a good job in not bottlenecking performance. But more testing is need and faster GPUs.

Optimal FCLK vs MCLK from Buildzoid [google spreadsheet link]
Attachment 285552


LAB501 3900X Memory Review does include 1440p and 2160p results for Ashes of the Singularity, Batman Arkham Knight, Dawn of War 3, Deus EX Mankind Divided, Metro Exodus, FarCry New Dawn, Shadow of the Tomb Raider, Shadow of War
https://lab501.ro/procesoare-chipset...nglish-version
Attached Thumbnails
Click image for larger version

Name:	Optimal FCLK vs MCLK from Buildzoid.png
Views:	49
Size:	20.7 KB
ID:	291486  


Overclocking History (Click to show)
Intel 386DX/16 @ 20 MHz
CX486SLC/50 @ 50 MHz
AM486DX2/80 @ 100 MHz
AM486 DX4-100 @ 120 MHz
AK5x86-P75 [email protected] 160 MHz
Pentium 90 @ 120 Mhz
IBM 6x86L P166+ [email protected] 160 MHz
Celeron 300a @ 454 MHz
Pentium 3 700 @ 933 MHz
Athlon XP 1400+ @ 1600 MHz
Athlon XP 1800+ @ 2088 MHz
Athlon XP 2500+ @ 2200 MHz
Athlon 64 3000+ @ 2603 MHz
Athlon X2 BE2300 @ 1900 MHz
Athlon X2 4200+ @ 2210 MHz
Athlon X2 5000+ @ 2600 MHz
Core 2 Duo E4300 @ 3000 MHz
Core 2 Duo E6750 @ 3600 MHz
Core 2 Duo E8400 @ 4000 MHz
Xeon C2Q L5408 @ 3200 MHz
Xeon C2Q X5450 @ 3300 MHz
Xeon C2Q E5440 @ 3400 MHz
Xeon C2Q E5450 @ 3600 MHz
Phenom2 X4 B50 @ 3700 MHz
Phenom2 X6 1045T @ 4050 MHz
Core i5-2500K Sandy @ 4500 MHz
X5650 6-Core Westmere @ 3982 MHz
AMD FX 8320E Black Edition

FX-8320E @4.5GHz, X5650 @ 4.0GHz, X6 1045T @4.0GHz, i7 [email protected]
Thinkpad T420p with i7-2720QM overclocked because this is OCN
Heuchler is offline