Message boards :
Number crunching :
Radeon R9 Cards Not Passing POST
Message board moderation
Author | Message |
---|---|
Wedge009 Send message Joined: 3 Apr 99 Posts: 451 Credit: 431,396,357 RAC: 553 |
I'm hoping some of you might be able to help me here - I've already done a lot of searching around the 'net in general but haven't found any solutions yet. I have two R9 280Xs and one R9 270 (all happen to be Gigabyte brand). I've verified the R9 280Xs work on a new motherboard with PCIe 3.x support. The R9 270 is also running fine, on a motherboard with only PCIe 2.x support. However, when I try putting the R9 280Xs (whether one at a time or both at once) in an older but working motherboard, also only supporting PCIe 2.x, I see no POST messages on start-up - only a black screen. The monitor does register a video signal, it's just completely blank. CPU and GPU fans all spinning and no warning lights. All this happens with the R9 270 as well. Older PCIe 2.x GPUs work just fine on this motherboard, and in the past I've also had a GTX 670 work okay. I also tried the R9 270 on an even older motherboard still (R9 280Xs are too big for the case), with only PCIe 1.x support. Again, no POST, and again, older GPUs work fine. All motherboards are running at latest BIOS. I understand that video cards and motherboards of different PCIe versions are supposed to be largely compatible with each other (albeit with corresponding performance penalty for older versions, of course). But it looks to me as though my older motherboards cannot handle the R9 generation (I don't know if the same thing would happen with HD 7000 series even if they are mostly the same). I have a recollection of testing a GTX 780 on the old motherboard as well, but dismissed the lack of POST because it was faulty on my new PCIe 3.x motherboard too. I've seen suggestions of switching outputs (VGA/DVI/HDMI/DP) but that doesn't appear to change anything. Obviously, BIOS updates aren't an option (which I've seen some people say was a solution for them). Is it possible that older BIOSs cannot handle the current generation of video cards? Why would this be the case? (I thought any incompatibilities would be limited to CPU support.) I can accept it might be possible that one cannot run a new card in an older computer, but would like to know why as I consider my options (don't want to have bought the 280Xs for nothing). Thanks for any thoughts, ideas or suggestions you can offer. Soli Deo Gloria |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13855 Credit: 208,696,464 RAC: 304 |
Have you tried swapping power supply units from the systems that do run with the cards to the ones that won't? Have you got power to all the connectors (both the 8 & 6 pin)? During boot the load should be bugger all, but I've had stranger things happen. Grant Darwin NT |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Thanks for any thoughts, ideas or suggestions you can offer. Did you plug in the extra motherboard power connector (usually toward the backplate connectors) ? "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Wedge009 Send message Joined: 3 Apr 99 Posts: 451 Credit: 431,396,357 RAC: 553 |
Sorry, yes, forgot to mention that I've plugged in all extra power connections and made sure they're secure. On the motherboard and the video cards. Soli Deo Gloria |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Sorry, yes, forgot to mention that I've plugged in all extra power connections and made sure they're secure. On the motherboard and the video cards. I figured. just eliminating the obvious. what about the case ? do its expansion slots allow the card to fully seat without gymnastics ? "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13855 Credit: 208,696,464 RAC: 304 |
I'd check AMD's site & see if there is a BIOS update for the video cards. It's possible the older systems expect a particular response during POST, if they don't have a BIOS update to remedy that, the video card manufacturer may. Grant Darwin NT |
Terror Australis Send message Joined: 14 Feb 04 Posts: 1817 Credit: 262,693,308 RAC: 44 |
Once again stating the bleeding obvious, but have you checked the video settings in the motherboard BIOS ? I had a similar problem some time ago that was caused by the BIOS being set to look for a PCI video card (the default) on boot up instead of a PCIE (or PEG) card. If you plug an older card that works into the motherboard, along with one of the R9's. What happens ? Do you get video out of the older card and does it show that the R9 has been recognised by the mobo and the OS ? T.A. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13855 Credit: 208,696,464 RAC: 304 |
I had a similar problem some time ago that was caused by the BIOS being set to look for a PCI video card (the default) on boot up instead of a PCIE (or PEG) card. Good point. It shouldn't make any difference, but with the newer hardware they may have neglected allowing for such older video arrangements, and the system just hangs there in limbo as a result. Grant Darwin NT |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
I had a similar problem some time ago that was caused by the BIOS being set to look for a PCI video card (the default) on boot up instead of a PCIE (or PEG) card. Hmmm, took me a while to think about, and possibly not related. My 680 and 480 wouldn't work on my old asrock 'crossover' board at all. It had both AGP and PCIe (rare), and works fine with old cards. None newer than some point. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
chromespringer Send message Joined: 3 Dec 05 Posts: 296 Credit: 55,183,482 RAC: 0 |
Sorry, yes, forgot to mention that I've plugged in all extra power connections and made sure they're secure. On the motherboard and the video cards. I'm throwing this out there as a shot in the dark ..... I had a similar issue where the machine would not post or even spin the fans for more than a second .. during the process of elimination I found that an OCZ power connector worked great in an OCZ power supply, however, for what ever reason would not seat tightly in an XFX power supply even though it "appeared" to make a secure connection possibly the 8-pin or 6-pin connectors are not actually seating correctly in the card for what ever reason .. try different connectors assuming your using a modular power supply |
Wedge009 Send message Joined: 3 Apr 99 Posts: 451 Credit: 431,396,357 RAC: 553 |
what about the case ? do its expansion slots allow the card to fully seat without gymnastics ? Yes, everything is seated properly - the motherboard shows a warning light if the card is not properly inserted into the PCIe slot. I'd check AMD's site & see if there is a BIOS update for the video cards. It's possible the older systems expect a particular response during POST, if they don't have a BIOS update to remedy that, the video card manufacturer may. I haven't found anything in this regard, with AMD or Gigabyte. Do you know where I should be looking? I had a similar problem some time ago that was caused by the BIOS being set to look for a PCI video card (the default) on boot up instead of a PCIE (or PEG) card. I re-checked the BIOS settings and confirmed PEG/PCI is the order used (as opposed to PCI/PEG). If you plug an older card that works into the motherboard, along with one of the R9's. What happens ? Do you get video out of the older card and does it show that the R9 has been recognised by the mobo and the OS ? I thought about trying this before. As it turns out, even Windows XP picks up the 'new hardware'. Without updated drivers (which drops OpenCL support in WinXP - my intention was to use the new cards under Linux), GPU-Z can't read the R9 280X, but it's recognised in CPU-Z (albeit as HD 7970). So this hardware combination does appear to function, but it just can't get through POST when used as primary video card. My 680 and 480 wouldn't work on my old asrock 'crossover' board at all. It had both AGP and PCIe (rare), and works fine with old cards. None newer than some point. Yeah, given the evidence I've seen so far, the newer cards will only work in older motherboards up to a point. The newest card the motherboard could handle was a GTX 670 - but not a GTX 780. On AMD side, HD 6000 series is fine, but looks like HD 7000 / R9 200 is not. I had a similar issue where the machine would not post or even spin the fans for more than a second .. during the process of elimination I found that an OCZ power connector worked great in an OCZ power supply, however, for what ever reason would not seat tightly in an XFX power supply even though it "appeared" to make a secure connection All fans are spinning upon power-on, plus I am not mixing cables and power supplies - I'm using the cables that came with the PSU and they work with the older video cards, just not the new R9 cards. Well, thanks for all the suggestions. Looks like I'll have to re-think my hardware arrangements. Soli Deo Gloria |
Terror Australis Send message Joined: 14 Feb 04 Posts: 1817 Credit: 262,693,308 RAC: 44 |
Just as a matter of interest. What is the motherboard/processor combination on the rig that's giving you grief ? T.A. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
The change in the PCIe spec from 2.0 to 2.1 causes some newer cards, of PCIe 2.1 or newer, not to work correctly in older motherboards. The motherboard manufacture may have an update to aid in this comparability issue. If not you may need to find boards that have PCIe 2.1 or 3.0 for the newer cards to be fully functional. The other option does seem to be placing a compatible card in the system. Maybe you could get a PCIe x1 card to boot the system if you have a free slot for it. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Wedge009 Send message Joined: 3 Apr 99 Posts: 451 Credit: 431,396,357 RAC: 553 |
ASUS P5E3 with Core 2 Q9550. Actually, it's this host. I thought HD 6000 series was PCIe 2.1... which works fine. Also, GTX 670 is PCIe 3.0. Soli Deo Gloria |
Terror Australis Send message Joined: 14 Feb 04 Posts: 1817 Credit: 262,693,308 RAC: 44 |
ASUS P5E3 with Core 2 Q9550. Actually, it's this host. This may sound crazy, but did you try booting up with a monitor connected to each card ? There was a quirk in some ASUS P5 BIOS's that when 2 PCIE GPU's were installed, the motherboard would get confused as to which one to boot to and would boot to GPU-1 instead of GPU-0. Maybe in your case the R9's have triggered it, even if the bug doesn't appear with other cards. T.A. |
Wedge009 Send message Joined: 3 Apr 99 Posts: 451 Credit: 431,396,357 RAC: 553 |
It's using 2 PCIe cards now, and has always shown boot screen from GPU0. Also, when I had the two R9 280Xs in it, there was only a video signal received from GPU0 - nothing from GPU1. At least, the monitor on the second card was in 'power saving' mode, which I take to mean no video signal is received. Soli Deo Gloria |
Terror Australis Send message Joined: 14 Feb 04 Posts: 1817 Credit: 262,693,308 RAC: 44 |
Scratch that idea then :) T.A. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
ASUS P5E3 with Core 2 Q9550. Actually, it's this host. Starting with the HD5000 series they were using PCIe 2.1. MY HD6870 worked fine in my C2D system as well. I just recall reading about this issues as I was thinking about getting some current generation low powered cards for my older systems. Something like the R7 240 or R7 250X. The change in the spec has to do with power from the slot & the implementation of the change could vary from device I suppose. It just seems like it may be your issue, or it may not. Computers can be a pain like that I found. I had a system that would not post if memory was inserted in channels A & B, or just A for that matter, & would give a "no memory present" post code. One of the screws holding the heatsink mount to the motherboard was "to tight". Despite using my torque screwdriver to tighten them all the same amount. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
chromespringer Send message Joined: 3 Dec 05 Posts: 296 Credit: 55,183,482 RAC: 0 |
just out of curiosity, what AMD driver are you using |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13855 Credit: 208,696,464 RAC: 304 |
just out of curiosity, what AMD driver are you using No driver, it won't even boot. Grant Darwin NT |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.