Overclock.net › Forums › Graphics Cards › AMD/ATI › Polaris Bios Editing ( RX5xx / RX4xx )
New Posts  All Forums:Forum Nav:

Polaris Bios Editing ( RX5xx / RX4xx ) - Page 147

post #1461 of 3348
Guys, ATOM BIOS reads for the PowerPlay on a 470 the following:

000f: 9c9a Len 032c Rev 07:01 (PowerPlayInfo)

But WHERE exactly is located this? Searching for that 9c9a , gave me exatly 2 results, but i dont know which is the correct one, and where are the starts/ends. I am looking into disabling ROPs
Desktop PC
(21 items)
 
ASUS ROG G72GX
(6 items)
 
 
CPUMotherboardGraphicsGraphics
AMD FX-8320 ASRock Fatal1ty 990FX Killer Sapphire Nitro+ RX480 XFX RX470 Singlefan 
RAMHard DriveHard DriveHard Drive
Mushkin Redline 996996 2x4GB 2133Mhz Maxtor 6Y080L0 80GB 7200 RPM 8MB Western Digital 160GB 7200RPM 8MB Maxtor 250GB 7200RPM 8MB 
Hard DriveHard DriveOptical DriveCooling
Corsair Force LS WesternDigital Blue 500GB 7200RPM 16MB ASUS DVD-RW ThermalTake Frio Silent 14 
OSOSOSMonitor
Windows 10 Enterprise Linux Mint 17.3 Rosa OphCrack ASUS VS228HR 
KeyboardPowerCaseMouse
Logitech K120 Corsair VS650 Delux Bloody V5 
Audio
Corsair HS30 Raptor 
CPUMotherboardGraphicsRAM
Intel Mobile Core 2 Duo P8700 G72GX NVIDIA GeForce GTX 260M  Hyundai Electronics  
RAMRAM
Hyundai Electronics  Hyundai Electronics  
  hide details  
Reply
Desktop PC
(21 items)
 
ASUS ROG G72GX
(6 items)
 
 
CPUMotherboardGraphicsGraphics
AMD FX-8320 ASRock Fatal1ty 990FX Killer Sapphire Nitro+ RX480 XFX RX470 Singlefan 
RAMHard DriveHard DriveHard Drive
Mushkin Redline 996996 2x4GB 2133Mhz Maxtor 6Y080L0 80GB 7200 RPM 8MB Western Digital 160GB 7200RPM 8MB Maxtor 250GB 7200RPM 8MB 
Hard DriveHard DriveOptical DriveCooling
Corsair Force LS WesternDigital Blue 500GB 7200RPM 16MB ASUS DVD-RW ThermalTake Frio Silent 14 
OSOSOSMonitor
Windows 10 Enterprise Linux Mint 17.3 Rosa OphCrack ASUS VS228HR 
KeyboardPowerCaseMouse
Logitech K120 Corsair VS650 Delux Bloody V5 
Audio
Corsair HS30 Raptor 
CPUMotherboardGraphicsRAM
Intel Mobile Core 2 Duo P8700 G72GX NVIDIA GeForce GTX 260M  Hyundai Electronics  
RAMRAM
Hyundai Electronics  Hyundai Electronics  
  hide details  
Reply
post #1462 of 3348
So, there's a thing I think I noticed about ram overclock and modded timings stability.
Everything was ok when I had my old case, max temp. was 55c and the card was perfectly stable. Now I changed case (I still have set some things up and add a couple of fans) and the card hits 66-67c after a long usage.

I see lightning artifacts, little purple flashes, I think it's typical in these cards when something's unstable.
It never happens when I'm around 60c. The problem is that it happens very seldom, like one time per hour and it's so fast I probably miss some of them.

The doubt that it may be core instability has come to me, but I have been using 1075mv/1300mhz for some months now and I thought it was pretty solid.

Argh, I can't sort it out. Is it a temperature "issue" or am I fooling myself?
post #1463 of 3348
Hello i have RX470 Sapphire 8gb Nitro with Micron memory. Can someone help to flash it please, if you can target me where to download rom file:rolleyes:
post #1464 of 3348
Quote:
Originally Posted by daevy3k View Post

So, there's a thing I think I noticed about ram overclock and modded timings stability.
Everything was ok when I had my old case, max temp. was 55c and the card was perfectly stable. Now I changed case (I still have set some things up and add a couple of fans) and the card hits 66-67c after a long usage.

I see lightning artifacts, little purple flashes, I think it's typical in these cards when something's unstable.
It never happens when I'm around 60c. The problem is that it happens very seldom, like one time per hour and it's so fast I probably miss some of them.

The doubt that it may be core instability has come to me, but I have been using 1075mv/1300mhz for some months now and I thought it was pretty solid.

Argh, I can't sort it out. Is it a temperature "issue" or am I fooling myself?

I'm convinced that most overclocks that people deem stable are not. In the 7950 world, one driver update messed up people's stability, and they said the driver was bad. Meanwhile my oc still worked with it... They were too far on the edge of stability for a measly 5%...

For cpu use Intel burn in test, linx, prime95. For gpu, only ones that really tell you numbers of errors are oclmemtest (looped) and the best OCCT!

Also, unless you have some really crazy cooling, like n2, make sure the overclock is stable at high temperatures that are accepted by oem standards.

That's why I run my i5-2500k At 4.5, not 4.7 because I wanted it to be stable in case the cooler loses efficiency. Same for my 480 at 1275, because any more and occt gives errors and random crashes
Edited by robnitro - 3/11/17 at 11:19pm
post #1465 of 3348
Thread Starter 
Quote:
Originally Posted by ku4eto View Post

Guys, ATOM BIOS reads for the PowerPlay on a 470 the following:

000f: 9c9a Len 032c Rev 07:01 (PowerPlayInfo)

But WHERE exactly is located this? Searching for that 9c9a , gave me exatly 2 results, but i dont know which is the correct one, and where are the starts/ends. I am looking into disabling ROPs

9c9a is the stating point. 032c is the lenght
post #1466 of 3348
Thread Starter 
Quote:
Originally Posted by stenli560 View Post

Hello i have RX470 Sapphire 8gb Nitro with Micron memory. Can someone help to flash it please, if you can target me where to download rom file:rolleyes:
Post your bios here. I'll help. What do you want to change in you bios?
post #1467 of 3348
Thread Starter 
Quote:
Originally Posted by robnitro View Post

I'm convinced that most overclocks that people deem stable are not. In the 7950 world, one driver update messed up people's stability, and they said the driver was bad. Meanwhile my oc still worked with it... They were too far on the edge of stability for a measly 5%...

For cpu use Intel burn in test, linx, prime95. For gpu, only ones that really tell you numbers of errors are oclmemtest (looped) and the best OCCT!

Also, unless you have some really crazy cooling, like n2, make sure the overclock is stable at high temperatures that are accepted by oem standards.

That's why I run my i5-2500k At 4.5, not 4.7 because I wanted it to be stable in case the cooler loses efficiency. Same for my 480 at 1275, because any more and occt gives errors and random crashes

I don't really like using occt on cards with a boost feature.(nvdia 600+ or amd 7000+) This cards are not built for an extreme load like that. At stock "power limit" a throttle will occur, and by incressing the power limit + occt you are not testing the real world stability(also you are over-stressing the power section). I prefer to play 2-3 games that stress the card at 100% and than up the voltage by one step and i will be 99% stable. A game will never cause a vdrop/stress like occt.
Our cards are literally built for throttling when a extreme load is applied.
Thats my opinion.

Cpus are different, they are very sensitive to errors.
Edited by generaleramon - 3/11/17 at 11:57pm
post #1468 of 3348
Quote:
Originally Posted by generaleramon View Post

I don't really like using occt on cards with a boost feature.(nvdia 600+ or amd 7000+) This cards are not built for an extreme load like that. At stock "power limit" a throttle will occur, and by incressing the power limit + occt you are not testing the real world stability(also you are over-stressing the power section). I prefer to play 2-3 games that stress the card at 100% and than up the voltage by one step and i will be 99% stable. A game will never cause a vdrop/stress like occt.
Our cards are literally built for throttling when a extreme load is applied.
Thats my opinion.

Cpus are different, they are very sensitive to errors.

Yeah with boost, I do a bit of a trick to test a more steady game rate:
I find the max fps loaded, let's say 400 fps.
Then I try 300 fps and watch the clock, it should be higher and throttle less.
Powertune tuning also helps with occt. Let's say you want to find out the max steady power that doesn't overheat your card... you back down on powertune until it's stable and not overheating. For example on my rx480, around 150 watts is the max steady reliable power.. past that GPU can go above 80 C and start throttling power on temperature.

I run my occt test at 640x480 with fps set to where I use around 130 watts for a steady test with the powertune around 150 watts (calculated from my bios set to 140watt, using 7% pt to get 150 total watts where card throttles).

Games don't tell me whether there are errors or not, once you see an artifact, there's been hundreds/thousands of errors that lead to that!

The rx480 is less crazy, but on my 7950, if I didn't do my occt torture testing for error free, it would randomly give me issues in some games, depending on the level. Some levels stress memory or the shaders more and you won't know until you're deep into the stage and boom, crash or some really bad artifacting!
post #1469 of 3348
Quote:
Originally Posted by robnitro View Post

I'm convinced that most overclocks that people deem stable are not. In the 7950 world, one driver update messed up people's stability, and they said the driver was bad. Meanwhile my oc still worked with it... They were too far on the edge of stability for a measly 5%...

For cpu use Intel burn in test, linx, prime95. For gpu, only ones that really tell you numbers of errors are oclmemtest (looped) and the best OCCT!

Also, unless you have some really crazy cooling, like n2, make sure the overclock is stable at high temperatures that are accepted by oem standards.

That's why I run my i5-2500k At 4.5, not 4.7 because I wanted it to be stable in case the cooler loses efficiency. Same for my 480 at 1275, because any more and occt gives errors and random crashes

Yeah, but what is considered to be a good result in ocl memtest?
Results seem pretty unpredictable to me, usually I get around 15kk errors, but some times I get around 3-6kk errors without apparent reason. If I lower memory controller voltage I get lower number of errors but it just happens until reboot even if I keep settings.

You are right about oc margin, but I'm running my RAM at 2000, while I can reach 2150mhz without apparent problems.

Just one error every 10-15min in hwinfo and very rare flashing artifacts. It's hard to say when it is really stable :/
post #1470 of 3348
Yeah, there's too many variables. On my 7950, I was pulling out my hair on oclmembench errors and hwinfo errors. My memory was at stock and still got errors! It turned out to be that the gpu overclock was sometimes messing up the memory, probably it's own IMC was causing errors???

A lot of things linked together can make issues that show up. Just like CPU overclocking or tuning a car engine, the weak part shows up when you least expect it!
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD/ATI
Overclock.net › Forums › Graphics Cards › AMD/ATI › Polaris Bios Editing ( RX5xx / RX4xx )