In polaris bios editor, where do you undervolt the gpu, for exemple to 900mv ?
ROM Vgpu changes just ignored (reason unknown). Just make downvolt with Sapphire Trixx
Not all of them, just the ones doable with PBE.
Sure, I know. Base voltage for 300 MHz gpuclock is changeable, I mentioned it in this topic. I'm increasing it from 800 mV to 875 or 900. It helps to keep card stability during undervolting with Trixx.
So there's no way to edit the bios to apply static mV drop (to all voltage table) without having to use Trixx?
There is - for Sapphire Nitro+ cards, all the ones I've seen have one more Voltage ID (and value) in the voltage regulator init stuff over stock, and it's the one you need you set a global core offset that will work on all platforms. By default, its value is set to 4 - the way this value works is that it is a signed byte - times 6.25mV. So they overvolt the core (all states) by 25mV. You could change it to -25mV by setting it to 0xFC.
If it's not a Nitro+, you'll need to add the value, and that's major surgery to the VBIOS.
i change the driver. now i can mine ETH, but cannot mine ZEC. where is the problem? any help much welcome.thx
Both ZEC and ETH miners have some undefined issues with starting right after using some 3D apps (games, etc), not everytime, but sometimes. Can not find dependency.
But everyting OK after system reboot. Maybe it's something drivers or OpenCL libs related.
I have observed also something strange : after few days, sometimes, a fan jump over 90%, on my 6 Rx480 rig. Everything is OK after a reboot : 60° / 60%
1125/2080 getting 17mh 1125/2000 getting 26-27mh used silent_nitro1625to2000strap+VDROP any suggestion?
1. Check page 5 or 6 of this topic for last V4 ROM mod. 2. Unlock voltage regulation in MSI AB 3. Your memchips are not so lucky to keep 2080 MHz 4. You use dual mining mode, 29+ Mh is only for single ETH mode.
I've been lucky with one card and unlucky with another. very good bios when it works! Is a similar bios released for the nitro rx470 8gb ? I cannot find a good one around here
My rig config is 6xSapphire RX 480 8G modded bios with powersafe rom H81 PRo BTC Win 10 64bit 16.9.1 Crimson driver USB powered risers
I have a strange problem with Claymore Ethereum miner 7.4 (ZEC miner is OK). The matter is that the hashrate of GPU0 is 14-16 MHs (all the rest 5 cards are 29MHs) GPU0 is the card that connected to 16x PCI-E slot and the monitor is connected to it. I tried to switch the "main" card with another 2 cards(reconnect the riser to 16x PCI-E slot) and the result is the same after reboot (GPU0 shows 14-16MHs, GPU1-GPU5 show 29MHs). So the problem is not in the card. Maybe somebody faced somewhen with such problem? Are there any solutions of it?
My rig config is 6xSapphire RX 480 8G modded bios with powersafe rom H81 PRo BTC Win 10 64bit 16.9.1 Crimson driver USB powered risers
I have a strange problem with Claymore Ethereum miner 7.4 (ZEC miner is OK). The matter is that the hashrate of GPU0 is 14-16 MHs (all the rest 5 cards are 29MHs) GPU0 is the card that connected to 16x PCI-E slot and the monitor is connected to it. I tried to switch the "main" card with another 2 cards(reconnect the riser to 16x PCI-E slot) and the result is the same after reboot (GPU0 shows 14-16MHs, GPU1-GPU5 show 29MHs). So the problem is not in the card. Maybe somebody faced somewhen with such problem? Are there any solutions of it?
My rig config is 6xSapphire RX 480 8G modded bios with powersafe rom H81 PRo BTC Win 10 64bit 16.9.1 Crimson driver USB powered risers
I have a strange problem with Claymore Ethereum miner 7.4 (ZEC miner is OK). The matter is that the hashrate of GPU0 is 14-16 MHs (all the rest 5 cards are 29MHs) GPU0 is the card that connected to 16x PCI-E slot and the monitor is connected to it. I tried to switch the "main" card with another 2 cards(reconnect the riser to 16x PCI-E slot) and the result is the same after reboot (GPU0 shows 14-16MHs, GPU1-GPU5 show 29MHs). So the problem is not in the card. Maybe somebody faced somewhen with such problem? Are there any solutions of it?
Thanks for help.
Try to test this card alone.
As I said before I tried several cards set as "main"(connect to 16x PCI-e slot and plug monitor to it) and always GPU0 shows 14-16 MHs, GPU1-GPU5 29MHs. The matter is not in specific card.
The only way I could get any GPU to run in the "main" PCIE slot was setting it's memory clock lower, 2035 for a sapphire RX8g480, 2080 for XFX GTRs. On risers all cards run 2170MHz / 30.5 Mhs . All gpus have moded bioses. If I i raise the memclock I just get artifacts and your usual BSOD. Anyone got an ideea why cards seem more unstable in the PCI-E slot then on risers? I also tried placing another 480 in the boards second PCI-E slot, same result.
Can anybody think of a reason why with Jukebox's original and V4 BIOSes I can "only" get to 26Mhs instead of the 29+ everybody else seem to get? I am on ETHOS with sg_miner Many thanks
Can anybody think of a reason why with Jukebox's original and V4 BIOSes I can "only" get to 26Mhs instead of the 29+ everybody else seem to get? I am on ETHOS with sg_miner Many thanks
Hello, new miner here ;-). I just made half of my rig (still waiting for 3 additional cards). Been mining for a day using v4 bios (big thanks for that) and noticed a strange thing, don't know if it's normal or not, my hashrate drops after few jobs by 3-20MH. I am using Claymore dual miner, win10, 16.9.2 drivers, 3xRX480 nitro+ 8GB.
@blasteris , try to adjust gpu and memory clocks/voltages, each gpu has it's own sweetspot. For monitoring hashrate the best thing you can do is find stable settings and see what your avg hashrate will be over the next 2-3 days.
Comments
I'm increasing it from 800 mV to 875 or 900. It helps to keep card stability during undervolting with Trixx.
If you leave base voltage unchanged (0.8 V), you'll have only 0.7 Vgpu when idle. It is too little to keep stability.
If it's not a Nitro+, you'll need to add the value, and that's major surgery to the VBIOS.
If I change the Gpu :
@1375 : 177
@1340 : 175
@1300 : 172
And same sol rate with mem @2000 or 2100 MHz
Jukebox, catch it
But everyting OK after system reboot. Maybe it's something drivers or OpenCL libs related.
1125/2080 getting 17mh
1125/2000 getting 26-27mh
used silent_nitro1625to2000strap+VDROP
any suggestion?
2. Unlock voltage regulation in MSI AB
3. Your memchips are not so lucky to keep 2080 MHz
4. You use dual mining mode, 29+ Mh is only for single ETH mode.
Is a similar bios released for the nitro rx470 8gb ? I cannot find a good one around here
My rig config is
6xSapphire RX 480 8G modded bios with powersafe rom
H81 PRo BTC
Win 10 64bit
16.9.1 Crimson driver
USB powered risers
I have a strange problem with Claymore Ethereum miner 7.4 (ZEC miner is OK). The matter is that the hashrate of GPU0 is 14-16 MHs (all the rest 5 cards are 29MHs)
GPU0 is the card that connected to 16x PCI-E slot and the monitor is connected to it. I tried to switch the "main" card with another 2 cards(reconnect the riser to 16x PCI-E slot) and the result is the same after reboot (GPU0 shows 14-16MHs, GPU1-GPU5 show 29MHs). So the problem is not in the card. Maybe somebody faced somewhen with such problem? Are there any solutions of it?
Thanks for help.
Anyone got an ideea why cards seem more unstable in the PCI-E slot then on risers? I also tried placing another 480 in the boards second PCI-E slot, same result.
Many thanks
I am using Claymore dual miner, win10, 16.9.2 drivers, 3xRX480 nitro+ 8GB.