x264.exe Crashing with new Mobo

[Crono141]

Hi all.

I had a motherboard failure, and had to replace it with a Gigabyte GA=MA785GM-US2H. Running Phenom II quad core 2.24 ghz, 4 gb of ram, using HDC as my encoder gui. x264 consistently crashes at around 10% completion on a 1 pass crf quality encode. I've tried different builds, and they all crash. I just finished reinstalling windows because I thought it might have been caused by a driver issue between the old board and this one, and still no luck. Is there a known issue with my new hardware and this Mobo?

More importantly, what can I do to help fix it? Thanks.

[poisondeathray]

hardware issues? bad ram? temps? psu?

check your temps, rum memtest, prime95, linx , occt etc...

did you try another source? does it crash on that too?

[Blue_MiSfit]

No known issues AFAIK.

Are you certain your source file isn't corrupted? How does x264 crash - does it literally crash with an error, or does it just hang?

I'm almost totally sure it's a problem with your hardware ;). Run the usual stability test (prime95, memtest, etc...) Also, give us your full system specs.

This probably belongs in the PC Hardware and Software forum, to be honest. Maybe a mod will kindly move it for you :)

~MiSfit

[blubberbirne]

Hi all.

I had a motherboard failure, and had to replace it with a Gigabyte GA=MA785GM-US2H. Running Phenom II quad core 2.24 ghz, 4 gb of ram, using HDC as my encoder gui. x264 consistently crashes at around 10% completion on a 1 pass crf quality encode. I've tried different builds, and they all crash. I just finished reinstalling windows because I thought it might have been caused by a driver issue between the old board and this one, and still no luck. Is there a known issue with my new hardware and this Mobo?

More importantly, what can I do to help fix it? Thanks.

i had the same problem with my Core i7 system a few month ago. The RAM Voltage was to low.

[Blue_MiSfit]

Could be that. Memtest would / should reveal a problem there :)

~MiSfit

[burfadel]

Is it common across different encoding files? what is the source for the encoding? If its just with one particular file its the source, if its common across many different sources then its something else (of course). Since you said 10 percent crash it sounds like you are comparing the same source file with different builds? If it were memory timings/voltage, it would be a lot more random than consistently at 10 percent! The other possibility is there is something non-dodgy about the source that still causes a bug to appear in x264 (most likely) or a filter, but this is much less likely than the source being at fault.

[Crono141]

Hi all.

Thanks for the good advice, and pointing me to the stress testing tools. With my original configuration, I had all 4 stress tests fail in under a minute, usually also causing a BSOD. Got Speedfan to monitor temperatures, and core heat was 122 F when things started crapping out.

I opened up the case, moved some HDD around, put a case fan in front of the HDD, moved a CPU fan to exhaust processor heat, and opened up a CD/DVD slot in the case to allow cooler air to flow in from the front of the PC. I also moved the PC from the floor in an enclosed space onto a shelf into a more open space. 15 minutes into this encode, core temp hovers around 117 F, what I can only assume is memory and chipset temps hovering at around 127F and 131F, and ambient case temp at 99F. No crashes so far, but it did indeed seem heat related.

Yeah, this should go into Hardware. I just didn't realize it at the time of creation that it was a hardware problem.

:thanks:

[poisondeathray]

130F is about 54C . It's very unlikely that would cause a crash from overheating.

BSOD errors are usually memory issues , or graphics card driver , or BIOS related

I would try to rule out memory first, run memtest from USB boot

[Blue_MiSfit]

Do make sure your CPU heatsink is properly attached. 122f (50c) is pretty low to have problems - seems a bit odd.

Check your memory timings and voltages - make sure the memory is getting adequate voltage per it's spec, and that the timings match the spec for the speed you're running.

~MiSfit

[LoRd_MuldeR]

Check your memory timings and voltages - make sure the memory is getting adequate voltage per it's spec, and that the timings match the spec for the speed you're running.

Indeed. While memory should run fine at standard voltage with standard timings, some "low latency" memory modules may require higher voltage when running with reduced timings.

My memory modules were giving my BSOD's too, after I reduced the timings from the standard values to the vendor's specifications.

I had to increase the voltage - according to the vendor's specifications - and now all works perfectly fine. That said, be careful when raising the voltage, too high voltage can cause damage!

Last but not least you should run Memtest86+ for several hours to check you RAM modules. If you see errors, then either the modules are bad or not compatible with your mainboard.

It's recommended to get RAM modules that are listed on the compatibility list of your mainboard. Most vendor's offer such lists...

[jakor]

I had core 2 duo, one of the first ones, 8200 running at 90 C for several months and no problems detected.

[Crono141]

Seems the problem is more than just heat. While I can get through encodes now, I woke up this morning discovering 2 of my HDDs were reporting SMART errors.

Fun...:mad:

[BigDid]

Hi,

If you have not changed your power supply, have you checked that it can handle your new rig electrical needs?

Did

[Crono141]

Well, I can`t figure out whats up. Tried messing with voltages, and that made things worse. Memtest showed no errors. Running the short fft test in Prime95 runs indefinitely, Blend tests and long tests crap out in under 5 minutes consistently. x264 still crashes or causes freezes when run. Not sure what else to do. Got a new 650W power supply, so thats not the issue.

[burfadel]

First, go in to the bios and load bios defaults. Unless you are overclocking the default voltages should be ample and setting them high can just cause heat and/or longevity issues.

Try this:
Download cputweaker from: http://tweakers.fr/beta.html

Don't change the CAS Latency or the Command Rate, since you can't really change those in Windows sucessfully, but for the rest of them select slower timings. For Refresh Cycle Time for example, choose a much slower settings (say if its 50 now change it to 80+), for the RAS precharge change it from say 5 (or whatever it is now) to 8 and do the same for the rest of them.

After applying, try encoding etc agan. Even though it is a tweaking programme (you generally choose faster settings with it), you can go the other way :) Make note of where the settings are at the moment, then if it works you can try resetting each setting back to its original value until you find the one causing the problem. You can then find the fastest workable setting then either set it permanently in the bios, or save the settings and set the cputweaker settings automatically each time you boot :) You may be lucky and have to set one setting slower whilst being able to tweak other settings.

Boosting voltages doesn't overcome some timing issues, its actually much better, especially if you haven't overclocked, to find out why at the default voltage and speed the (for example) the current RAM settings don't work like they should!

[Crono141]

Ran that program, and noticed something weird right off the bat. CPU core frequency kept changing from x11 to x5.5, and back again, for no apparent reason. About once a second.

EDIT: Ok, tried playing with timings using this program. Any changes I made either made things worse or BSOD'd. This also tells me that memory is a factor in my failures though, otherwise I would think that changing these settings wouldn't really matter one way or the other. Maybe I also don't know exactly what I'm doing either.

I got another SMART failure on my drives. I run (2x) 1 TB drives and a 500 GB drive as a dynamic JBOD disk (through windows disk manager). When I get the smart failures, its always on the 2x 1 TB drives. I usually get these errors when running a stress test and the computer freezes (no BSOD, just frozen). Can failing HDDs cause my Prime95/x264 issues, or is this a separate problem?

Rundown of the hardware attached. 7-zipped html file.