Overclock.net banner
Cancel
Create thread New Forums

GIGABYTE AORUS X670E MASTER Discussion thread

Tags
aorus x670e
Jump to Latest
201 - 220 of 956 Posts
#201 ·
GRABibus said:
Yes, come on AMD DDR5 OC thread and read :LOL:

I mean, Karhu's is the test to pass at 1:2 on 2DPC boards.
All others are easy.

And Karhu's, you can pass 30000%, then at next boot, you fail at 100%.
Typical fail on 2DPC boards at 1:2.

@domdtxdissar @wmdieselmc @BlueShark @Yuke @Eusebiugh and all others can confirm ;)
Click to expand...
You're the RAM police, then ? Anyway, I'm using Intel XMP microcode (no EXPO) RAM; besides, I have been building systems, including servers, for a few decades now and am quite happy with my long RAM stress tests expertise. So let's leave it at that...

DedEmbryonicCe1 said:
The ECLK on this board is really handy for a quick and dirty OC to benchmark with. I'll have to go back later and find stable settings, but I was able to do much better in my 3dmark sessions the last day. I think unstable CPU causes memory errors even in ASYNC mode though. I tried tightening my tRFC to what my A-die should be capable of and Memtest would run for a while clean then suddenly pop up a big batch of errors all on one core then continue chugging along without further errors. I was just too excited with my new GPU to take the time to find something 100% stable before I started benchmarking. Currently at -30 CO and 101.90 MHz BCLK and it's 99% stable. I'm getting tons of driver crashes from pushing my GPU too hard but that's what you got to do to be on top for a while.
Click to expand...
...certainly agree on eclk being useful and even fun on this board, but aware of false RAM speed readings in HWInfo, CPUID etc...just posted on that this morning over >here
 
#202 · (Edited)
J7SC said:
You're the RAM police, then ?
Click to expand...
What a stupid comment.

We all know that you built rigs for decades and then have nothing to learn.

maybe my message was not so friendly and appears like I was giving you some advises, but that’s was not the case. I was just mentionning what has been raised concerning stability on 2DPC mobos at 1:2
 
  • Rep+
Reactions: Puneet724
Save
Reply Quote Rep+ Rep+
#203 ·
J7SC said:
...certainly agree on eclk being useful and even fun on this board, but aware of false RAM speed readings in HWInfo, CPUID etc...just posted on that this morning over >here
Click to expand...
Well that would explain why I was failing MemTest with a ECLK OC and now that I removed it I'm passing. I'll have to do more testing to see where I can go for 24/7. I was just trying for max OC for 3dmark and it did well for that purposes.
 
#204 · (Edited)
I'm kinda sure that the weird memclock shown in the bios is just a graphical glitch, not the actual memclock the system is running.

I'm currently running eclk 105, hynix 24gb m die *4 at 6000-30-36-36, and absolutely stable.
Which, if the memclock shown on the bios was true, It would mean that I should be running ~6300-30-36-36, 96gb, at around soc 1.23, vddio 1.25, vdd 1.38.

I tried exactly the same voltage to see if 6400-30-36-36, or 6200-30-36-36 would be possible without eclk, with bclk=100, but, naturally, it would not even boot.
Actually, since my ****ty memory controller cannot even run 24*4 6200-32-37-37 fully stable, I am certain that, this is the evidence that even with ECLK 105, my memory clock is actually running at the proper rate: 6000-30-36-36, following actual bclk, instead of following whatever random number the bios is throwing up.

I suspect that probably, just gigabyte being gigabyte, decided to outputs whatever bclk or eclk times multiplier, on the bios screen, causing this suspected graphical glitch we are all experiencing.


...On the other hand, I have no evidence, but suspecting that my bclk or eclk itself is somhow 1) fluctuating very severely, and 2) not actual 100 bclk, or 105 eclk, according to my input, or 100.8 bclk, or 105.8 eclk, as the bios screen shows, but somewhere between 'whatever number inputted + 0.4'.
 
#205 ·
Cpoint said:
I'm kinda sure that the weird memclock shown in the bios is just a graphical glitch, not the actual memclock the system is running.

I'm currently running eclk 105, hynix 24gb m die *4 at 6000-30-36-36, and absolutely stable.
Which, if the memclock shown on the bios was true, It would mean that I should be running ~6300-30-36-36, 96gb, at around soc 1.23, vddio 1.25, vdd 1.38.

I tried exactly the same voltage to see if 6400-30-36-36, or 6200-30-36-36 would be possible without eclk, with bclk=100, but, naturally, it would not even boot.
Actually, since my ****ty memory controller cannot even run 24*4 6200-32-37-37 fully stable, I am certain that, this is the evidence that even with ECLK 105, my memory clock is actually running at the proper rate: 6000-30-36-36, following actual bclk, instead of following whatever random number the bios is throwing up.

I suspect that probably, just gigabyte being gigabyte, decided to outputs whatever bclk or eclk times multiplier, on the bios screen, causing this suspected graphical glitch we are all experiencing.


...On the other hand, I have no evidence, but suspecting that my bclk or eclk itself is somhow 1) fluctuating very severely, and 2) not actual 100 bclk, or 105 eclk, according to my input, or 100.8 bclk, or 105.8 eclk, as the bios screen shows, but somewhere between 'whatever number inputted + 0.4'.
Click to expand...
...you are probably right; just had a similar discussion in another thread on that. However, I'm still trying to figure out the relative jump in SuperPi performance re. center entry (which btw is repeatable), given the relatively small bus clock steps involved. 3DVCache effect ? Error correction ? I haven't figured that out yet.

In any case, I have come to appreciate this specific CPU IMC / IF / RAM and mobo combo; pretty resilient ! Gigabyte's eclk documentation is a nothing-burger so far though when it gets to useful details, never mind that bios memory 'live reading'. And that 'clear CMOS 'jumper'...I had to visit several times now given that I was trying to find some limits...I keep that big-tip screw driver and flash-light within reach at all times 🤪

Colorfulness Font Rectangle Screenshot Parallel
 
#207 ·
Product Font Software Screenshot Technology


I dropped the memory from 6000 to 5800 so it wouldn't be a factor in this test. At BCLK 101.9 MHz and -30 CO no cores are failing so I'm free to go higher. This raised my max frequency from stock 5550 MHz to 5675 MHz.


Cpoint said:
I'm kinda sure that the weird memclock shown in the bios is just a graphical glitch, not the actual memclock the system is running.

I'm currently running eclk 105, hynix 24gb m die *4 at 6000-30-36-36, and absolutely stable.
Which, if the memclock shown on the bios was true, It would mean that I should be running ~6300-30-36-36, 96gb, at around soc 1.23, vddio 1.25, vdd 1.38.

I tried exactly the same voltage to see if 6400-30-36-36, or 6200-30-36-36 would be possible without eclk, with bclk=100, but, naturally, it would not even boot.
Actually, since my ****ty memory controller cannot even run 24*4 6200-32-37-37 fully stable, I am certain that, this is the evidence that even with ECLK 105, my memory clock is actually running at the proper rate: 6000-30-36-36, following actual bclk, instead of following whatever random number the bios is throwing up.

I suspect that probably, just gigabyte being gigabyte, decided to outputs whatever bclk or eclk times multiplier, on the bios screen, causing this suspected graphical glitch we are all experiencing.


...On the other hand, I have no evidence, but suspecting that my bclk or eclk itself is somhow 1) fluctuating very severely, and 2) not actual 100 bclk, or 105 eclk, according to my input, or 100.8 bclk, or 105.8 eclk, as the bios screen shows, but somewhere between 'whatever number inputted + 0.4'.
Click to expand...
It's hard to say because I definitely had MemTest errors with BCLK when I left my memory at 6000. Without any BCLK OC I pass MemTest fine so I don't know what else is being overclocked but something is influencing my tests.
 
#208 ·
Font Rectangle Parallel Screenshot Software
Colorfulness Font Screenshot Computer Software


This is what I'm getting if I try using ASYNC and BCLK. First screenshot from last night with no BCLK OC and second screenshot with 101.7MHz. It's making my memory unstable even though the cores pass prime95 in corecycler. I don't understand what the problem is yet but I'll keep testing.
 
#211 · (Edited)
By now, I've had enough time to really get into this new setup and very much appreciate the Aorus 670E Master, including on price : performance ! The performance collage below summarizes things pretty well. My daily DDR5 setting is 8000, though in the bottom right corner, you see a CL30 6200 (didn't try lower CL).That means A LOT OF flexibility. FYI, the top CPUZ was with Process Lasso locking on to CCD1, while the bottom one is the daily (both CCDs).

The memory topography in particular stands out re. this board, along with its VRM which isn't even breathing hard w/ a 7950X3D. The quirkiness re. clearing CMOS jumpers isn't the best solution (nor its labelling), but I don't get to use that much anymore :D. I also have a few different profiles for eclk tuning, but I like to work some more on those just for the fun of it before I share those...

Photograph Light Computer Font Screenshot
 
#212 ·
What does a stable system seven segment display, display? Currently I have 00 but it mostly displays 76.

It doesn't say what it should be when your system is running stable also does anyone know what 76 means when it is displayed?
 
Save
Reply Quote Rep+ Rep+
#214 · (Edited)
Hi Guys,
Any help would be very appreciated now. I have tried so many things to get my RAM working But still getting C5.
I have the trident z5 neo rgb (16 x 2).
My Specs :

Motherboard : Aorus Master X670E
RAM Type : F5-6000J303816GX2-TZ5NR
DDR5-6000 CL30-38-38-96.
CPU : Amd Ryzen 7800X3D

On the first day, both of the RAM were working fine, but with 4800, I tried to enable XMP and do some custom changes. Both were working fine as well. I installed windows, tested my 4090, all was working fine until the restart. After restart I got my first c5, I flashed all BIOS versions available, reset CMOS, etc. Still c5, i then removed both of the RAM's and started testing them individually, now It booted with one of the RAM's. The other RAM was giving C5 on all the slots. Today, I messed with some XMP profiles by setting VDD and VQQ by some recommendations in this thread and now both of my RAM's are not working. Please note they stop working only after restarting the computer from windows. Now I am out of options to get them working. Quite frustrated at this point.
One Thing keeps bugging me, I have reset CMOS but still the rgb settings are persisted, where are the rgb settings stored. My GPU, AIO, CPU fans all remember the rgb patterns from RGB fusion application. So I am thinking maybe my RAM settings are stored there too and resetting them in some way would bring my RAM's back.
 
#216 ·
lucian2pinro said:
Hello !
Does anyone saved the 10a bios ?
I need that bios version.
TY
Click to expand...
...don't know if this is 10 a specifically, but it is the only '10' still listed > here

Elsewhere, having a bit of fun with DDR5 8200 :) - using the stock 7200 / daily 8000 timings and don't think these DDR5 8200 are 'bench stable', but it is still a pleasure to overclock the RAM on this Aorus 670E with its memory topography...

Font Screenshot Parallel Software Multimedia
 
Discussion starter · #217 ·
J7SC said:
...don't know if this is 10 a specifically, but it is the only '10' still listed > here

try here think it has all released bios

www.aorus.com

'X670E AORUS MASTER｜AORUS - GIGABYTE United Kingdom'

'AMD Socket AM5：Supports AMD Ryzen™ 7000 / Ryzen™ 8000 / Ryzen™ 9000 Series ProcessorsUnparalleled Performance：Twin 16*+2+2 Phases Digital VRM SolutionDual Channel DDR5：4*SMD DIMMs with AMD EXPO™ & Intel® XMP Memory Module SupportNext Generation Storage：2*PCIe 5.0 x4 and 2*PCIe 4.0 x4 M.2...
www.aorus.com www.aorus.com

slow and full of java crap may have to reload the page a couple of times 10, 10a, 10b, etc are on there along with 27 more (30 in total including the iffy ones )
Click to expand...
 
201 - 220 of 956 Posts
About this Discussion
955 Replies
99 Participants
Last post:
PhantomR
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling