Jump to content

Ethrem

Registered User
  • Posts

    352
  • Joined

  • Last visited

  • Days Won

    2

Everything posted by Ethrem

  1. Hi Prema, I have a question. Is the XTU + modded nVidia vbios = no POST a Clevo issue or is it something to Prema Mod? I had 32_32 on my SM-A and I had to revert from the modded vbios back to stock because my system would get stuck when I would would hit reboot where it would show the modded vbios screen and then shut itself off and back on and keep doing that until I pulled the power from the machine, let it shut off, then pushed the power button and booted it from being powered off before plugging in my power cord. D2 Ultima had the same issue with his machine and 780M. Modded vbios would cause the system not to POST when XTU was used. Its either have XTU or have the modded vbios but not both. Today I had 32_33 and it was working fine on full EFI but when I switched the BIOS to non-EFI so I could boot to flash another image, after I turned off EFI it would not POST. I took out the battery, reset the CMOS battery... I ended up having to take one of my cards and stick it in my old Alienware and flash with the stock vbios to get the system to POST again. I wasn't having any issues with booting or shutting down but I had johnksss help me today and he was changing XTU settings and gave me a new vbios to flash and the reboot worked on EFI but when I changed it to non-EFI to do the flash, that's when it bricked. Any idea of what the cause is? Should I just not use legacy BIOS and stick with UEFI when using a modded vbios? Some insight here would be great because I'm totally confused. :|
  2. Bricks happen. I almost had to send my machine off for repair after the initial 880M vbios release had an error that caused it to brick machines without Optimus, things happen. I believe svl7 is on vacation, at least that was the last I heard. Maybe Klem or someone else here can shine light on what happened and your vbios questions. Sent from my HTC One_M8 using Tapatalk
  3. I'm amazed they even offered to reflash it for you. Don't flash the vbios if you can't accept the risks associated with it. Things don't always work how they are expected to. Clevo machines won't POST with a bad vbios flash so you have no option but to either put it in another MXM machine that has on board video and doesn't link vbios to bios (Alienware basically) or you can buy vbios chips on eBay and solder them yourself to get back to stock vbios or you can send the card off to get reflashed like you did. It's interesting in your case though because it did POST. Curious did you turn off boot logo in BIOS settings? I did that the other day and then couldn't get into the BIOS either. I had to boot to Windows successfully and shut down (not restart) then turn the machine on and mash the F2 as soon as the back light came on to get in there, then had to reset defaults and reboot and now it works fine again. Seems to be a bug. Sent from my HTC One_M8 using Tapatalk
  4. Thanks for all the work you do Prema! Sent from my HTC One_M8 using Tapatalk
  5. Does this mean that there is hope they don't EOL P377SM-A and they release an official bios update? Because that would mean I can upgrade through Xotic and have a warranty on the new cards. Sent from my HTC One_M8 using Tapatalk
  6. I want that 980M for my P377SM-A! Sent from my HTC One_M8 using Tapatalk
  7. Its the cycle.. New things come out, the old things go away. If it wasn't that way, people like myself wouldn't have gotten stuck with the trash that is 880M. It doesn't matter if the new product isn't better than the old product, Intel, nVidia, and AMD have seen to it that the old part supplies dry up when the new parts come out. Sent from my HTC One_M8 using Tapatalk
  8. Is your power supply light going out when this happens? That would tell you if you're tripping over current protection. My guess is that you are. Sent from my HTC One_M8 using Tapatalk
  9. Did your set on boot setting get disabled? I don't know why everyone loves that program, nothing but problems for me and less fine tune voltage adjustment than Precision X has. I always got annoyed when I would set one card, go to set the other, and come back to the main card just to see the settings didn't take and have to do them again. Personally I've not had a great experience with MSI or Inspector but I have zero issues with the old Precision X (I don't install RTSS though because I have that installed for use with Hwinfo). Sent from my HTC One_M8 using Tapatalk
  10. Sager had 20 left the other day. Clevo announced that they are killing the machine. They are likely working on a new X99 Haswell-E machine or they just decided to kill it. Either way, it's end of life and Clevo isn't known for supporting their products with updates after that. Sent from my HTC One_M8 using Tapatalk
  11. Good luck with that seeing how they just discontinued that machine. Sent from my HTC One_M8 using Tapatalk
  12. That's the default behavior. You need a program to set the custom clocks on boot. I use Precision X to do it personally. In order to have your overclock apply all the time without software setting it, you would have to modify the vbios with your own custom values. That's what makes overclocking safe - if you reboot the machine or it freezes, shuts itself off, etc, the values are cleared. Sent from my HTC One_M8 using Tapatalk
  13. Yes it forces boost clocks all the time. Sent from my HTC One_M8 using Tapatalk
  14. I don't consider it something to brag about but a card performing at its advertised clocks is less broken than a card that doesn't perform at its advertised clocks and overheats when you force it to. Imo anyway. Sent from my HTC One_M8 using Tapatalk
  15. I missed that generation entirely but did they at least hit their stock clocks out of the box? Sent from my HTC One_M8 using Tapatalk
  16. Yeah man the 880M really is the worst release nVidia ever made. I run mine @ 849MHz @ 0.896v so I don't have to do anything stupid to keep temps under control. Max temps with full fans is 72C, max fan with auto is 84C and the fans never actually have to spin up because they don't hit 87C which is when Clevo fans go to their high state so my machine is much quieter now. It's a sacrifice but all my games run flawlessly with it but it's basically now a 780M with 8GB of RAM. You shouldn't have to make such a sacrifice on a card that costs so much money. Sent from my HTC One_M8 using Tapatalk
  17. I don't buy 70C for a non-throttling 880M. Have you logged this with GPU-Z and confirmed it's not throttling? If so, who makes your machine and what's your stock vbios version? I'm not aware of any stock vbios that doesn't throttle. The only exception *may* be the 4GB ASUS card based on benchmarks I've seen. Sent from my HTC One_M8 using Tapatalk
  18. Gaming on battery is going to a challenge and attempting it with cards that don't down clock can kill your battery as in physically damage it. As far as safe voltages the cards will hit their thermal shutdown point long before you would do electrical damage to them. 1.1v is likely to overload a single A/C adapter though. My system can't do 1.1v on both cards even with turbo turned off on my 4940MX. Fire strike will pass but 3dmark 11 will overload it. Sent from my HTC One_M8 using Tapatalk
  19. So I'm back on the modded vbios. Apparently there is a bug with XTU, Clevo BIOS, and the mod where using XTU can cause the reboot issue I was describing. D2 Ultima on NBR alerted me to it. I removed XTU and flashed the mod and it is now working 954MHz @ 0.981v set on boot with EVGA Precision X. Testing it now but since the cards are stable with the stock vbios and it runs the cards at that voltage for that core, I think it shouldn't have any issues. Sent from my HTC One_M8 using Tapatalk
  20. CPU-Z should show your model in the motherboard screen. The BIOS also tells you I believe. Sent from my HTC One_M8 on Tapatalk
  21. I just had to flash it back to stock. For some reason whenever my machine would restart, it would pop up with the modded vbios message then shut off, turn itself back in and repeat that cycle until I pulled the power cord from the back. It would shut down one last time then I had to turn it on with no power cord or it would loop again. Now that I flashed them back to stock, everything is fine. I am at a loss as to why they would start to behave like that. I also could not pass the physics test on 3dmark cloud gate. Both 337.88 and 340.56 drivers crashed and recovered 4 times each time the test initialized. I tried bumping the voltage up to 1.025v to no avail. Just ran the test on the stock vbios and it completed with no issue. Any ideas here? Man I hate these cards... Sent from my HTC One_M8 on Tapatalk
  22. MSI cards need a modified version of the nVidia drivers I thought? http://laptopvideo2go.com/drivers/win8x64/337.88/disclaimer Get the driver and the INF from there, extract the driver, and replace the installer INF with the modded one and try it. It's just a guess but it's worth a try. Sent from my HTC One_M8 on Tapatalk
  23. Yeah the only way I actually stress these cards out is by throwing AA settings I can't even notice at them, even just one is more than adequate for most games. The two combined are faster than my desktop 780 Ti which only gets a slow down with ultra textures in Watch Dogs because of only having 3GB of RAM. Here's what happened with Firestrike all options maxed out in custom (including 8xMSAA, 16xAF, all tesselation and other options, 1080p), looped on first scene for an hour and sixteen minutes Max temp was 90C with auto fans, I turned the fans on max at the end which was good for 1-2 degrees off, honestly don't think its worth the added noise. So worst case scenario for a game, probably looking at 88C on auto, 85-86C on max because the load in a game is almost never near 100% while the 3DMark run is extremely close as you can see. I'm very happy with this result but of course, it leaves zero thermal headroom for overclocking later.
  24. I didn't try with this set but my last pair couldn't do +20 on stock voltage. Highest core I got on the old ones was 1098 @ 1.1v which was able to pass firestrike - http://www.3dmark.com/fs/2326858 Unfortunately I only had one 330W A/C adapter at the time so both 3DMark 11 and Catzilla overloaded it. Since these actually work, I'm not so sure I want to risk overclocking them. Sent from my HTC One_M8 on Tapatalk EDIT: Actually, I got them to 1153 - http://www.3dmark.com/fs/2327316 I had to turn off turbo for the CPU for that test so the score is a bit lower than it should have been.
  25. Yeah I don't plan to overclock at all. I just wanted my cards to stop dropping the slave to as low as 888MHz for no good reason, I was even willing to set the base clock to 954MHz instead of 993MHz if the heat was too high but to my surprise, everything seems to be fine. Bioshock Infinite does burn up the cards but I can turn on adaptive vsync and that won't be an issue anymore. Sleeping Dogs is kind of the same, just have to disable extreme AA which I can't really tell apart from high anyway and temps are down a few degrees. All in all, I'm finally pleased. I think it was bad cards from the start, at least that slave, because Sager put the same heatsink on the new slave and its obviously not even hitting 90C now. The best I was able to get last night was -12.5mv, -25 crashed the first card in Valley almost immediately and it took about 10 minutes before the slave crashed with it. As you can guess, 12.5mv didn't really change the temperature at all. Congrats on getting the temps down, I'm a big fan of that paste. Sent from my HTC One_M8 on Tapatalk
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.