Discussion in 'TechPowerUp Club Forum' started by the54thvoid, Mar 19, 2017.
Do you like the board
I have the same one
I so far like the size of the board and it being simple. Got this board mostly because of the price being the lowest, the size and I just want something simple. This board almost in my itx case. Just that it is 1 slot to high. It would be perfect if it 3 slots high only.
What is your cpu clocks for this board?
I haven't finished my rig as I haven't had time
So not sure how high my 1700 can go
Awesome bundle but I still feel the lack of features on the Ryzen boards that's r on the Intel boards like on-board Creative Core3D.
I got the Gigabyte Aorus Z270 Gaming 7 with Creative Core3D the Ryzen X370 Gaming K7 r using the Realtek ALC1220 chip with SoundBlaster X-Fi MB5 it just isn't the same
Update on my new BIOS.
Running 109.2 BCLK as the standard DOCP was flaky. Memory seems stable now at rated SPD/case of 3200/14.
Currently on Prime all cores at 3.8Ghz (35x). temps stable at 70 degrees after half hour. Don't want to press luck but it might be more stable now.
Thinking of maybe trying to lower latency (rather than hghr frequency on mem or core) but have no idea what I'd be doing....
yeah, not a great board TBH
had a better experience on biostars
FWIW , ladies and gents who have the giga x370 gaming k7 rev 1.0 they've put up bios F5 for release.
Fix P2-state hanged issue while changing CPU Vcore of Ryzen 3
Improve DDR compatibility
Other giga am4 boards may have also some new updates.
As far as I've read, they included more dividers that was added for agesa 1006a and added an option for linux users to disable opcache function.
Started experimenting with Beam.NG to see how far I could tax the CPU, but ended-up crashing the game by hitting max vRAM first
Just wanted to share something regarding gigabytes x370 K7 BIOS or maybe other board BIOS that they released on 2017/08/18.
K7 owner here and reporting some findings about the mysterious F5 bios that came out on the official site (it didn't go thru the BETA bios thread), it was just posted as a "stable" release.
Having flashed the bios some users noticed it to set the cpu vcore to a very high voltage 1.55v instead of the default 1.2ish~ v. Having applied optimized settings saved and restarted it went back to normal.
Findings: Most or some users got the same experience.
More info, it is something to do with auto and voltage offsetting that this bug occurs, but if you set it to use a manual voltage and use LLC to compensate then you are golden.
At the moment have read and found out some people getting their processors fried due to this bug. We haven't heard yet from gigabyte in their forums regarding this one.
Minus that very very bad bug, overall the bios is very good. It has unlocked more dividers/mhz for the memory (3066, etc.).
+Noticed with the same oc settings that I use, the VRMs are also way cooler than previous set of bioses also my machine idle temps are way lower (atm running sweet spot 3.8ghz on all cores)
+XMP on some kits even on other dies that the godly b-dies can hit XMP with just turning on the setting without further tweaks. (I had to raise ram voltages to make it stable tho 1.4v)
+The notorious soft brick issue (for me maybe) does not happen (haven't noticed it too on F4). Maybe I just go lucky.
I was not aware of the vcore bug as I always set my voltages manually, it was after a few hours of the bios being released that I found some user feedbacks regarding that new bios on the 18th made cpu vcore very high (some other flavor of the am4 gigabyte boards).
As of date I saw 2 or 3 persons on the gigabyte forum reporting of fried processors.
Just leaving it here for brothers and sisters who use a gigabyte board.
Sharing is caring
Dude, you made the news.
He made it here too.
I did scroll down the TPU front page to check before posting, but didn't see it as it's further down than you would expect.
Ohohoho what a way to wake up in the morning. Apparently giga took the bios down and replaced it with a fixed one.
That bios was really good but the vcore part made it undesirable.
Some can clock xmp speeds without tweaking. However my m-dies need a few tweaks with ram sub timings and voltages.
The VRM definitely is way more lower in temps.
(Now F6 or latest fixed bios carries the same properties minus the vcore bug).
Alas I hope those who got their processors damaged will get some compensation, but highly doubtful about that.
That's the annoying part. How do you prove that the CPU was damaged by that BIOS at those settings? It's just as possible that the CPU was killed in any number of other ways, so no way to prove it in practice and Gigabyte gets away with it.
Damn it.. *Shakes fist!*
Well further update regarding the newer bios, everything is in order.
I just hate that they didn't even made it as a beta bios and put it up on the site as a stable release and they didn't event say a word about it. Just put up hey new bios for those having issues with f5 et al.
They don't even have a beta thread anymore. They "accidentally" deleted their old one and "can't restore it."
Sometimes I wonder WTF GIGABYTE is thinking...
they still have! And there is were we continued other info.
3400mhz cl 14 @ 1.4v using a 102 bclk. Haven't updated to the latest bios yet.
I recognize that set up!!! Small world bro
Asus Prime B350M-A motherboard (w/latest 22.214.171.124b) + 1500X. Can only seem to choose 3500-3800. If I select 3900+ MHz, then it seems to run at minimum MHz despite showing 3900+. Is this because it isn't X370 platform?
When set to 3900, then I see the clock toggling between 1550 and 3900 in my OS. But I assure you, the performance is ABYSMAL like almost 3x slower than 3800.
Edit: Found similar cases on Reddit. Seems like cpu voltage slams some B350 motherboards to min clocks. Ah well, case solved, even if it's limited. 1.24v booted 3.9 but it's probably not stable, will just have to upgrade the board later if I want to play with it further.
The power draw of the CPU or the VRM not supplying enough current? If the board isn't designed for OC, it may be the PCB hardware isn't up to scratch. Could that possibly affect the frequency by throttling it when pushed?
Also - 3.9Ghx of Ryzen is a pretty decent standard. I cannot hold stable at 3.9Ghz on a 1700x with a X370 board (well, not if I want my memory at 3200Mhz).
It actually seems analog-ish. +0.06875 (1.30625) may or may not be stuck at 1550 for that boot, but +0.0625 (1.3v) seems to not stick at 1550. It's not throttling, it's plain stuck at min clock forever when it kicks in. Just [B350] product segmentation, but the implementation is a bit misleading, you don't notice immediately unless benchmarking or observant ...
Using 1.3v, I've got 4Ghz + 3200 CL14 testing now, finally! Seems to work. 4050 starts throwing ecc corrections in dmesg. Settled for plain 3900, seems like a safer choice. My emacs is now optimized for best performance ...
At 4Ghz, 1.262v throws a black screen eventually, and 1.268v worked for 12 hours, so I opted for 1.293v which is a notch under my board's 1.3v limit, that's as good as this cpu+board combo is gonna do.
Any advice on SOC voltage? It defaults to 1.1v at DDR4-3200 speeds, but it seems to work fine when soc manually set back to 0.831v?
Just got my new parts for my main rig upgrade a couple days ago.
This is what I've got stable so far.
R7 1700 @ 3.9GHz 1.35V
2x8GB G.skill Trident Z B-die @ 3466 14-14-14-34 1.35V
The cpu doesn't seem to want to do 4GHz though, tried up to 1.45V and it still crashes.
Still need to do some tweaking for 3600 14-14-14 on the memory.
Prime95 knocked out my 4GHz. Wound up with 3.9 in the end.
Separate names with a comma.