If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
PC erratic behaviour
I've been having problems with my homebuilt PC for a couple of weeks.
All the components have been replaced this year. The last to be changed were the RAM (32Gb Crucial) and the PSU. The problems seem to coincide with a BIOS update on the Asus Prime X570 PRO mb. I can't revert to to the earlier version because the mb doesn't recognise the BIOS file as valid, and I've tried several, the option to update from the internet is no longer offered, and when I try using the EZ update from the PC it crashes the PC - both problems have been reported over several years and Asus don't seem to know how to fix them. What is happening is that the PC takes a long time to load Windows. When I can access the desktop and run Task Manager something has been taking control of the cpu, utilisation is coming down from 100% but no individual process is doing more than it should be. In between times the PC is usuable but there are signs that it running slower than it should be as occasionally I can see things being written to the screen in a way that I have not seen for years. Usually it is instant. If I open, say, Outlook, and then, say, Edge, and switch from one to the other there is always a pause while one is relinquished in favour of the other. Nothing shows in the TM cpu (AMD Ryzen 5 3600) performance graph to account for this temporary loss of control, no sudden peak in cpu use. Another inconsistency is that whe I shut down the PC, or put it to sleep, it closes Windows down but not the PC, the fans continue running. Sometimes, not always, wake up or restart is accompanied by a crash, with a different cause most times. After a few days of this I did a complete reinstall, which seemed to improve things for a few hours, until the erractic behaviour resumed. I tested the ram with memtest and no errors were reported. Tried swapiing the sticks around and running with only one of them but it made no difference. An online malware scan only reported potential unwanted programmes in the recycle bin and registry. Trend Micro's HouseCall, which I've found to be reliable in the past, got stuck on listing autorun entries so I stopped it after an hour. In view of the random crashes I tried changing the power supply. No crashes so far, only been in for a few hours, but otherwise no change. I've updated the Nvidia graphics driver (PNY Nvidia Quadro P2000 card). Asus support said to return the mb to vendor (Amazon) for repair/replacement/refund. I'm reluctant to do that as I don't know for sure that it is the cause of the problem but I might have to. Anyone have any ideas of what else I might try, before I bite the bullet and do battle with Amazon over the returning the MB. (I've had it for more than three months so have to try and find a person to deal with.) Many thanks. Shout up if there's more information required. |
#2
|
|||
|
|||
PC erratic behaviour
On Thu, 27 Aug 2020 16:18:07 +0100, Peter Johnson
wrote: The problems seem to coincide with a BIOS update on the Asus Prime X570 PRO mb. I can't revert to to the earlier version because the mb doesn't recognise the BIOS file as valid, and I've tried several, the option to update from the internet is no longer offered, and when I try using the EZ update from the PC it crashes the PC - both problems have been reported over several years and Asus don't seem to know how to fix them. You've another, or should, if I'm to go on the provision an AM4 socket, rather ASUS, still supports it. Look-up, download HIRENs and make yourself a bootable CD. You'll have a blinking thing called a cursor.The BIOS and BIN associative versions also should be directly obtainable from ASUS (MD5 checksums for added veracity) or general types supportive hardware sites. Follow the instructions on how to Run an EXE file provided for BIOS Flash Utility, including the /? parameter, and ascertain explanations should the options provided be unclear before a selected BIN is loaded and Flashed to CMOS from DOS. A hell of a note when the MB contracts out for a BIOS chip, then to tell its customer base, as many effectively will -- This procedure may be a consequent result you assume at your own risk, that we do not advise a potentially broken MB because the Flash you instigated didn't take. If our MB is working for you, leave well enough alone. That is our recommended course, because it's better you don't call us on our dime if you broke it with a Flash. So, then it gets Flashed anyway. I've never had a bad one. I've never used anything but DOS. I haven't pressed my luck, to the best of my recollection, and Flashed when I didn't need to or had cause to revert to an older BIOS version. If the BIOS works that it works may often stay true. Another thing, probably unrelated, but are you checking your BIOS CPU temperature sensor readings regularly, whether related for software or on general principle during system anomalies? A cool-running AMD doesn't need to throttle itself into a passive state of massive unresponsiveness. Hibernation is the first thing I personally turn off. I'm always on. Binary operation sector restores, from another OS, are also nice. While its running nice the binary image is taken. When it's not being nice the taken image is restored. AB if not BA. Platform development, adding software or "tweaking it", occurs experimentally between the two binary events as images taken progress in complexity, dated to subsequent stable states of evolving configurations. Shut down the PS from the back of the case if you don't like the logic of whether MB reset/pwr pins perform adequately. Sometimes a new PS is not a perfect world, may not help, say when booting between two different operating systems, when the MB's network loses coherence with the modem and the back PS switch is the fastest thing. When its working it's mostly always working. Malware scans are another one in an ideal world of progressive building upon a safe software platform. I don't use them. I'm particular about not having to quarantine what I let in to copy from the WEB to my hard drives for further testing. Autoruns I edit myself. Video cards can often be installed from a bare essence of drivers, auto-detected, or through a suite of sometimes extraneous material included with a vidboard purchase. Sometimes just like soundchips. A prior operational state means breathe or move so much a hair I don't like -- and boom, I'll reset the system to restore it with the last uncontaminated image -- before Joe Schmoe's software scheme of things tried out for auditions. I run in a standalone complex (from self-contained program executables located apart, on their own discrete storage subsystems, mostly, apart from any integral claims to operational system imaging). Moderately above average to minimally advanced in an essential complexity backups are, nevertheless, overall to computers. |
#3
|
|||
|
|||
PC erratic behaviour
Peter Johnson wrote:
I've been having problems with my homebuilt PC for a couple of weeks. All the components have been replaced this year. The last to be changed were the RAM (32Gb Crucial) and the PSU. The problems seem to coincide with a BIOS update on the Asus Prime X570 PRO mb. I can't revert to to the earlier version because the mb doesn't recognise the BIOS file as valid, and I've tried several, the option to update from the internet is no longer offered, and when I try using the EZ update from the PC it crashes the PC - both problems have been reported over several years and Asus don't seem to know how to fix them. What is happening is that the PC takes a long time to load Windows. When I can access the desktop and run Task Manager something has been taking control of the cpu, utilisation is coming down from 100% but no individual process is doing more than it should be. In between times the PC is usuable but there are signs that it running slower than it should be as occasionally I can see things being written to the screen in a way that I have not seen for years. Usually it is instant. If I open, say, Outlook, and then, say, Edge, and switch from one to the other there is always a pause while one is relinquished in favour of the other. Nothing shows in the TM cpu (AMD Ryzen 5 3600) performance graph to account for this temporary loss of control, no sudden peak in cpu use. Another inconsistency is that whe I shut down the PC, or put it to sleep, it closes Windows down but not the PC, the fans continue running. Sometimes, not always, wake up or restart is accompanied by a crash, with a different cause most times. After a few days of this I did a complete reinstall, which seemed to improve things for a few hours, until the erractic behaviour resumed. I tested the ram with memtest and no errors were reported. Tried swapiing the sticks around and running with only one of them but it made no difference. An online malware scan only reported potential unwanted programmes in the recycle bin and registry. Trend Micro's HouseCall, which I've found to be reliable in the past, got stuck on listing autorun entries so I stopped it after an hour. In view of the random crashes I tried changing the power supply. No crashes so far, only been in for a few hours, but otherwise no change. I've updated the Nvidia graphics driver (PNY Nvidia Quadro P2000 card). Asus support said to return the mb to vendor (Amazon) for repair/replacement/refund. I'm reluctant to do that as I don't know for sure that it is the cause of the problem but I might have to. Anyone have any ideas of what else I might try, before I bite the bullet and do battle with Amazon over the returning the MB. (I've had it for more than three months so have to try and find a person to deal with.) Many thanks. Shout up if there's more information required. https://www.asus.com/us/Motherboards...HelpDesk_BIOS/ Version 2606 2020/08/17 19 MBytes Version 2407 2020/07/03 18.89 MBytes Version 2203 2020/06/22 17.47 MBytes Version 1407 2020/04/10 15.38 MBytes Version 1405 2019/11/26 15.36 MBytes Version 1404 2019/11/08 15.32 MBytes * You will not be able to downgrade your BIOS after updating to this BIOS version Version 1201 2019/09/12 15.24 MBytes Version 0808 2019/07/17 15.13 MBytes Version 7010 2019/07/05 15.13 MBytes Version 0804 2019/07/05 15.13 MBytes Version 0604 2019/07/04 15.11 MBytes Version 0602 2019/07/03 15.11 MBytes That means the boot block was updated in release 1404, and it could be preventing going backwards. Depending on the CPU Support table, there might be a revision that if you went back far enough, you'd prevent forward progress again. In essence going backwards and bricking the system. Do not try to fix this by selecting a version too old to be useful. https://www.asus.com/us/Motherboards.../HelpDesk_CPU/ Apparently there is a RAM training interval, something at Agesa level, that might be going wrong. Some users, their system seems fine, others are slow to start. ******* https://www.reddit.com/r/Amd/comment..._motherboards/ "Keep in mind flashing/updating the UEFI will reset all defined parameters and operating profiles. You will not be able to restore defined values by using a UEFI Profile as profiles are not interoperable between builds. It is advised you note or screenshot (F12) your values prior to flashing if they are complex. Upon completing a flash I would recommend "load UEFI defaults" and completing a full reboot prior to redefining prior values." Standard advice from legacy BIOS days :-) ******* It appears these boards have enough issues with BIOS level, that perhaps the next portion of this post is totally unnecessary. ******* OK, the next time you reinstall Windows 10, *do not* connect to the Internet. Not at all. The reason it's mis-behaving, is Windows 10 installs drivers for all the items it finds. Asus probably has some ACPI object in the BIOS, and Windows 10 Device Manager scans and finds this, and is downloading something which ties up the computer. I would be interested to see for how long it behaves (like, across reboots), for as long as the network cable is disconnected. One other thing you should know, is the OS does not have exclusive usage of the PC. There is a thing called SMM mode (System Management Mode). The machine can take a System Management Interrupt (SMI) and the BIOS gets to run code it wants to run. In the past, with an OS such as WIndows 7, you might run DPCLat to measure "hidden activity", such as SMM activity. Some BIOS, the interrupt handler for SMM had a long run time, and the OS can't do anything while this happens. The OS doesn't even know it's been disconnected for a hundred microseconds while this is going on. We can thank Intel for this "innovation". Asus uses this routine to adjust multi-phase power solutions, like 30 times a second it can change the number of phases used by VCore, between 2 phases, 4 phases, or 8 phases. There were a couple small chips in the power supply sections on the board, that were getting register value changes. Go into the BIOS and disable all the power features that use silicon on board. This would be an attempt to shorten the SMM interval. The first time users became aware of this sort of behavior, was the Asus iPanel product, which relied on SMM to cause the LED display on the iPanel to be updated. An engineer made a mistake, and generated a too-wide interrupt signal, using some sort of RC circuit, and the OS was being held hostage for too long as a result. DPCLat was the tool of choice, by audio workstation guys, to detect bad designs and return them to Amazon when they weren't fit for purpose. But in Windows 10, DPCLat no longer works, so you need an older OS. DPCLat measures the Delay Procedure Call Latency (response time for an interrupt handler Ring 3 runtime to be completed - they're queued up until processed). The longer that gets, the more likely it is that SMM is doing too much stuff in the background. "Red bars are bad" https://www.thesycon.de/eng/latency_check.shtml That's just a suggestion, if the BIOS level load defaults doesn't help. If you can find a replacement for DPCLat, then you won't need a different OS in order to test. DPCLat is *not* in Wikipedia. I just found this, and haven't tried it or vetted it for reputation. https://www.resplendence.com/latencymon Paul |
#4
|
|||
|
|||
PC erratic behaviour
On Thu, 27 Aug 2020 14:32:45 -0400, Flasherly
wrote: On Thu, 27 Aug 2020 16:18:07 +0100, Peter Johnson wrote: The problems seem to coincide with a BIOS update on the Asus Prime X570 PRO mb. I can't revert to to the earlier version because the mb doesn't recognise the BIOS file as valid, and I've tried several, the option to update from the internet is no longer offered, and when I try using the EZ update from the PC it crashes the PC - both problems have been reported over several years and Asus don't seem to know how to fix them. You've another, or should, if I'm to go on the provision an AM4 socket, rather ASUS, still supports it. Look-up, download HIRENs and make yourself a bootable CD. You'll have a blinking thing called a cursor.The BIOS and BIN associative versions also should be directly obtainable from ASUS (MD5 checksums for added veracity) or general types supportive hardware sites. Follow the instructions on how to Run an EXE file provided for BIOS Flash Utility, including the /? parameter, and ascertain explanations should the options provided be unclear before a selected BIN is loaded and Flashed to CMOS from DOS. A hell of a note when the MB contracts out for a BIOS chip, then to tell its customer base, as many effectively will -- This procedure may be a consequent result you assume at your own risk, that we do not advise a potentially broken MB because the Flash you instigated didn't take. If our MB is working for you, leave well enough alone. That is our recommended course, because it's better you don't call us on our dime if you broke it with a Flash. So, then it gets Flashed anyway. I've never had a bad one. I've never used anything but DOS. I haven't pressed my luck, to the best of my recollection, and Flashed when I didn't need to or had cause to revert to an older BIOS version. If the BIOS works that it works may often stay true. Another thing, probably unrelated, but are you checking your BIOS CPU temperature sensor readings regularly, whether related for software or on general principle during system anomalies? A cool-running AMD doesn't need to throttle itself into a passive state of massive unresponsiveness. Hibernation is the first thing I personally turn off. I'm always on. Binary operation sector restores, from another OS, are also nice. While its running nice the binary image is taken. When it's not being nice the taken image is restored. AB if not BA. Platform development, adding software or "tweaking it", occurs experimentally between the two binary events as images taken progress in complexity, dated to subsequent stable states of evolving configurations. Shut down the PS from the back of the case if you don't like the logic of whether MB reset/pwr pins perform adequately. Sometimes a new PS is not a perfect world, may not help, say when booting between two different operating systems, when the MB's network loses coherence with the modem and the back PS switch is the fastest thing. When its working it's mostly always working. Malware scans are another one in an ideal world of progressive building upon a safe software platform. I don't use them. I'm particular about not having to quarantine what I let in to copy from the WEB to my hard drives for further testing. Autoruns I edit myself. Video cards can often be installed from a bare essence of drivers, auto-detected, or through a suite of sometimes extraneous material included with a vidboard purchase. Sometimes just like soundchips. A prior operational state means breathe or move so much a hair I don't like -- and boom, I'll reset the system to restore it with the last uncontaminated image -- before Joe Schmoe's software scheme of things tried out for auditions. I run in a standalone complex (from self-contained program executables located apart, on their own discrete storage subsystems, mostly, apart from any integral claims to operational system imaging). Moderately above average to minimally advanced in an essential complexity backups are, nevertheless, overall to computers. Thanks for your comments. Without going into them line by line I agree with several of them. Probably I'm ignorant on the rest. I an interested in your comments on temperature though, because the start of my problems also coincided with a big increase in ambient temperature for several days, which I meant to mention but I forgot. I've never taken much interest in cpu temps normally because it's never been an issue. I don't think that I particularly stress the system in day to day use, either. However, when the problem started I removed the resistors from the cpu cooler fans and the case fans to see if it made any difference, which it didn't. The ambient temperatures have returned to normal and I've been running the PC with the case side removed, which has also made no difference to the operation. What I will try in the morning is removing the cooler and cpu and renewing the thermal paste in case the temperatures have caused a breakdown in the existing. I'll report back. |
#5
|
|||
|
|||
PC erratic behaviour
On Fri, 28 Aug 2020 20:11:21 +0100, Peter Johnson
wrote: Thanks for your comments. Without going into them line by line I agree with several of them. Probably I'm ignorant on the rest. I an interested in your comments on temperature though, because the start of my problems also coincided with a big increase in ambient temperature for several days, which I meant to mention but I forgot. Forgot, huh. Add a post to yourself, me or we do it all the time. I've never taken much interest in cpu temps normally because it's never been an issue. I don't think that I particularly stress the system in day to day use, either. However, when the problem started I removed the resistors from the cpu cooler fans and the case fans to see if it made any difference, which it didn't. The ambient temperatures have returned to normal and I've been running the PC with the case side removed, which has also made no difference to the operation. Neither did I. We share that, the case sides. Best case I've owned, an Antec, already cool for all aluminum Lanboy, I run without the side. Same for another generic case, which is laid on its (other) side, top open. Lanboy is an upright tower. What I will try in the morning is removing the cooler and cpu and renewing the thermal paste in case the temperatures have caused a breakdown in the existing. I'll report back. Resistors to the CPU cooler, that's a little unclear unless you mean you've some sort of slider switch with resistors for a speed-adjustable fan. These newer AMD CPUs, and the reason I first mentioned cooling is probably immaterial, is, it seems to me, that they're much more robust with their own AMD engineering for advancements in overheat safety measures. I'll relate why on the last of the Bulldozer AM3-Plus designs, although that's not to say there could be undesirable effects from a condition above temperatures, say, at reasonable expectation for a reasonable CPU-cooler setup. I've AMD3+, though, and not what you have -- the Ryzen platform I'd expect even better yet. Anyway, I stuck a SATA into the CPU cooler on mine, turned it back on, and ran for a day without CPU cooling on Bulldozer architecture before noticing. When I did take notice, to wonder why my computer up and decided to go nuts, I looked at the BIOS sensor readings -- the CPU was pegged at 100C for boiling. Did it stop it for a whole day, though -- not at all. Did it stop anything afterwards and were there further consequences -- evidently not. (It's a used AMD CPU pulled from a computer once running in South Korea I bought from an international parts dealer.) What I did notice before going to the BIOS readout was unacceptable throttling, as if a SATA controller ran smack into a wall of pending data sector reallocation units, on a plattered HDD, and was stuck in limbo. Or, as Paul mentioned in the ASUS site Hardware Forums, an engineer who dropped some bad handling routines, which ASUS shipped out, regarding memory. CPU cooling, like well cooling in general, is just normal to a computer powered on. Those memory issues, although I didn't read through the forum, just might concern me. Usually, I'd venture, there's a lot of people running your motherboard, but perhaps only some notice what you're noticing. Despite being a known and valid issue, why is that? Is their memory brand different, do they run less memory, is there anything different with CMOS/BIOS memory timings apart from your setup? ASUS, regardless, sent out that board. Meaning perhaps they've their own take on how to interpret potential memory latency lag, extraneous routines in utilizing memory. How much of an issue is it and how to circumvent its occurrence. And, besides ASUS forums, there's search engines for other good forums that do on occasion relatedly address a more popular MB, than any ol' no-namers, rebadged Dell MBs and such. That info he provided would seem a valid first clue on the wonderful world of trouble-shooting sometimes discernable issues. With you, sounds like your clock is ticking. Amazon may or not charge a 15% restocking fee. I'm not up on Jeff's standards for staffing and testing bad motherboards. If not, then ASUS is betting on solid satisfaction over any losses incurred, if in the case, some engineer dropped a paperclip into the design of your model. Could be decision time is around the next resolution corner. Tick-tock, tick-tock. Is it worth, possibly, swapping for another brand of memory and do you need more than 4G? Another MB altogether? That's another one I haven't dealt with but once or twice, an entirely base platform order -- MB, MEM, CPU/cooler -- that's sliding sideways, clear across the room, and out the window. Thank the gods. Just my take if I seem a little too pessimistic. (I'm _below_ a Ryzen, used parts, &etc.) |
#6
|
|||
|
|||
PC erratic behaviour
Update report.
I'm a bit further forward. Removed the cpu cooler and the cpu yesterday. The thermal grease did not look good. It had obviously been disturbed when I installed them and was not evenly applied. Removed and applied fresh, re-assembled and it's run ok since, with no random crashes since. I'm guessing this was aggravated by the increase in ambient temperature. I put the original power supply back in, too. But the system was still running slowly. I noticed that OneDrive was using a lot of cpu cycles so closed it down. (My data is stored on a NAS and some it is backed up to OneDrive so I don't know what it was doing, or trying to do, on the PC.) There was some slight improvement in operating speed and booting was quicker. Then I noticed that Windows Problem Reporting was also active at the top of the list of cpu use so I've turned that off and noticed more improvement. Ran MalwareBytes and no issues reported. The fan resistors that I mentioned are intended to slow the fans down, to make them quieter. Taking them out I haven't noticed any difference in fan noise. (And I can't tell you what difference they made to the speed.) There's still a delay in switching between programmes but nowhere near as bad as it was. As I type this task manager reports that the cpu is running from 5 to 15% capacity and the ram at 15% (with Outlook, Edge (three tabs) and Agent running) so I don't see any obvious reason for the delay. The OS is on a 500Gb NVMe SSD with 300Gb free so I don't see any explanation there, either. Windows has just shut down the ethernet connection with a 'Windows has stopped this device because it has reported problems. (Code 43)' message in the device manager. Removing it, deleting drivers, and scanning for new hardware has brought it back. We'll see what happens next. |
#7
|
|||
|
|||
PC erratic behaviour
On Thu, 27 Aug 2020 16:18:07 +0100, Peter Johnson
wrote: I've been having problems with my homebuilt PC for a couple of weeks. All the components have been replaced this year. The last to be changed were the RAM (32Gb Crucial) and the PSU. Just to say that I seem to be sorted. Thinking that there might be something amiss with the OS I did an in place Insider install yesterday - left it doing its stuff when I went to bed last night and it's been running as sweetly as it ever did all day today. Thanks to Flasherly and Paul for your inputs. (I'll leave the 'BIOS not recognised' for another day. I think that was a red herring.) |
#8
|
|||
|
|||
PC erratic behaviour
On 8/30/2020 7:59 AM, Peter Johnson wrote:
The fan resistors that I mentioned are intended to slow the fans down, to make them quieter. Taking them out I haven't noticed any difference in fan noise. (And I can't tell you what difference they made to the speed.) Why would you not just use the BIOS to control the fan speed so it can speed up if things get hot? It will run slow and quiet whenever heat is not a problem, and your processor is protected from the heat. Most Bios's allow significant variations in temp/speed adjustments. |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Erratic behaviour joystick port on soundblaster. | cyberfreak | Homebuilt PC's | 4 | August 22nd 08 04:04 AM |
Joystick Port Button erratic Behaviour | cyberdream | General | 7 | August 22nd 08 12:15 AM |
HP 3740 erratic | David Mooney | Printers | 2 | November 9th 05 02:13 PM |
Erratic booting | Captainstarball | General Hardware | 2 | August 9th 04 07:47 PM |
Erratic Mouse | lazya | Gateway Computers | 1 | January 28th 04 04:10 AM |