Hyper Transport sync flood error

24 replies [Last post]
dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

I have been getting this error randomly and it's mainly in games. The computer would lock up for 5 seconds then reboot and give me the error. It's been going on for maybe a month. The BIOS was updated already to the latest. I think it may have to do with overheating. It gets really hot in here. 81 degrees Fahrenheit. I also heard it may have to do with cpu voltages. I looked at the voltage through cpu-z and it jumps between 1.088-1.352

I have an intake fan installed to the side of the case but it's deactivated because it was loud when controlled by mobo. I may plug it back in but there are huge out flow fans at the top and back of case. I turned those up to medium speed.

EDIT: Also, sometimes when in game the sound breaks and I don't hear anything...when I log off it comes back. Happened 3 times the past week.

Starting to get concerned...

Motherboard: MSi 790xt-g45
CPU: AMD Athlon x64 5600+ 2.9 ghz
GPU: evga 250 gts
PSU: Rosewill Green series 700 watt 54 amps single 12v rail

chrisrice93
Offline
Joined: 05/20/2011
Posts: 59

It sounds like a problem with your ram to me
What voltage is your memory running at? When the system is under heavy load undervolted memory can cause it to crash.
You can do a memory error check by hitting delete when windows starts to load.
Do you mean its 81 in your room or case? If your cpu is 81 then thats pretty damn good but if its your room you can use realtemp to check the cpu temps. CPU temps should be at maximum 40C at idle and 75C under full load.
link for realtemp: http://www.techpowerup.com/realtemp/
Hope this helps :)

eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

If you are seeing 81°F, that is essentially room temperature. 81°C is getting hot, getting close to the danger zone.

If you are seeing huge shifts in voltages (and if CPU-Z is right that is a 30% shift) you will see some very strange behavior of the system. I'd recommend checking the PSU out to verify that it isn't feeding an unstable current to the mainboard. Bad voltages can screw up CPU and memory interfacing, and that is essentially what the Hyper Transport bus is for.

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

To clear up the 81 degrees thing, I meant that is how hot it gets in this room. I have a thing that takes the temp for the room.

Anyway, it did the error again when I opened up speedfan.

Here is picture of readings

With Speedfan, the minimum voltage is 1.08 and the max is 1.36. I really don't think this is normal at all. Unless this is what Cool'n'Quiet does?

cpusram.png
eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

Argh! Turn Cool'N'Quiet off, especially if you are overclocking!

It is a great feature, on systems that aren't running full throttle all the time. I use it regularly on mini-servers, as it helps to reduce their thermal footprint in cabinets. But for a game system, or rendering rig, or something that will be pushing hard, DON'T BOTHER!

Cool'N'Quiet reduces voltages and throttles the CPU down to, well, keep the system cool (Cool) and the fans slow (Quiet). Not a feature the typical gamer needs.

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

I turned it off and the voltages are much higher but I am still getting major fluctuation in voltages.

EDIT: Alright now I'm getting a much more steady voltage. I don't know if it's because I tampered with the settings in the BIOS or because I'm playing a game. This is what I'm getting playing Mass Effect with the 2nd picture

vcore.png vo.png
eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

That's better, but still wider than I would expect. Any improvement in games?

I'm running an overclocked Brisbane X2, so my voltages are higher, but I'm seeing around a .07 shift compared to up to .2 on your charts. I'm thinking that Rosewill is putting out crap. I like Rosewill gear, but not PSUs, hence my OCZ.

voltages-speedfan.jpg

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

Actually, I did extensive research on this type of psu before buying it. It is manufactured by ATNG which is supposed to be solid from what I have heard. This PSU series was reviewed by hardwaresecrets and they were surprised at how good it was. They gave it a very extensive test and it passed everything. 80 plus bronze certified and all. I have also heard Rosewill power supplies are hit or miss. This should definitely be a solid PSU.

I really don't think it's the Power supply. I think it's the motherboard because the Hyper Transport Sync flood error seems to be somewhat common with the 790xt-g45. I contacted MSi about it to see if they have a solution and no word yet.

The voltage on the CPU seemed to be steady and within .02 while playing games like the second picture in my previous post (you may not have seen it since I edited it in). It's .2+ while idle it seems, which is why I am confused. Seems like a mobo issue. Doesn't seem like a PSU issue at all.

After 4 hours or so of constant gameplay my PC turned off completely. I suspect it overheated. I plan to get a new CPU soon which will come with a better heatsink and all. I turned on the intake fan and my gpu is cooler which will then cause less heat to rise up. Hopefully that will help.

UPDATE: MSi replied and I got a response that had typos. A foreigner as usual. All he said was to switch out the RAM and take the CPU out and put it back in. Haven't gotten since I started this thread yet.
All the voltages are fluctuating like crazy now and I'm worried. I don't get it. Yesterday for a while it was like that second picture above, now it's back being crazy. So ridiculous.

eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

I've been doing a lot of reading, and ATNG (and Rosewill, as a relabeled part) are getting crappy overall scores. The "hit or miss" aspect, I'm sorry, but I think you got a miss!

For a lighter system, you probably wouldn't be seeing issues like this, but you are in a spot where the onboard VCG can't properly correct fast enough for changes on the 12V line, and that is where the changes in output current happen. I really recommend that you get (borrow) a voltmeter with live view (one that doesn't just show a ½-1 second average) and check the 12V line for variance.

The pull from your video card is probably causing "stability" by pushing a constant drain on the 12V line, keeping it from fluctuating, and making it easier for the mobo to generate the proper current. At low draw (GPU at low clock, e.g. at desktop, and CPU at low cycle) the PSU is generating way too much power and doesn't have EFI or PFC to correct overpower, and this is why you are seeing the fluctuation. This is also why I recommend to EVERYONE not to go crazy on wattage when you buy a PSU, as too much power can result in flaky (or burned out!) PSUs. Focus on 10-25% over the system's needed max wattage (room for upgrading), and look for multiple 12v rails (the CPU channel should be dedicated) and PFC.

From a hardware expert (and remember, THIS IS BASED ON PROVIDED INFO! I am not there, so there may be something additional that I can't add to the diagnosis) I recommend a PSU swap. I can't tell you how many times I've been called by a client who has an off the shelf system with similar behavior and I've swapped a $60 PSU into it and it suddenly calmed down and ran perfect.

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

I swapped out PSU's and I'm getting VERY similar readings. Cooler Master Elite 460 watt (18 amps on two 12v rails). I think something is up with the motherboard. It is reading that my 12V rail is getting 9.0 volts with both PSU's and when I look in to the BIOS it is around 12 volts. It bounces from 11.5 to 12.5 but it's not 9 volts like what Speedfan is saying. Something is out of whack here. I look at speedfan and it's not consistent with the BIOS readings with either PSU. BTW still have yet to have the error since the thread started.

Untitled.png
eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

Okeeeey, strong suspicions don't always make you right. VCU failure on the mobo most likely. Bummer.

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

Not sure what your first sentence meant. I was thinking it was the mobo rather than the PSU since I got different readings from the BIOS than software programs which were inconsistent. In fact I remember everything being pretty flat lined when I first looked at it. I played games and I never did see that reading above again that was within .02. I would rather it be the PSU than the mobo since RMA'ing that is a pain in the ass especially if I have to contact microsoft about deactivating the OS for that mobo (hardest thing to get out of the PC on top of that). Anyway, what does VCU stand for?

I may check the PSU's with a voltmeter to triple check but I don't know how to use one or have one.

eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

I was saying that what I suspected was wrong.

VCU is "voltage control unit", and it is something that rarely fails like this, on modern boards, anyway. It is a series of circuits that break down the inputs to provide the power the CPU needs. The current design includes PFC between layers, and most VCUs are three to seven layers deep, so there is a lot of wiggle room in error correction. You normally see a bad one just burn out.

Good luck.

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

That's kind of scary. What worries me most is the 9.8 volts the 12V rail is getting. That's not good at all. I could have sworn it was around 12 volts a few weeks ago. The BIOS reads it at around 12 volts. So do I trust the BIOS over the software? Is it possible the software is not working correctly? I would assume anything below 11 volts would cause an immediate crash and it has not crashed.

I don't know what to do at this point.

eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

Yeah, the BIOS is reading directly off of the thermistor, so trust it over the software. Any chance you may have it misconfigured and it is looking for a different chipset? I recently swapped over to Core Temp which is driverless (runs as a single file), can be downloaded as 32-bit or 64-bit, and reads temps off of the chipset driver than directly, as Speed Fan does, and therefore tends to crash the system less. I upgraded Speed Fan from 4.37 to 4.43 and it was BSOD all day long. Uninstall Speed Fan and give Core Temp a shot: http://www.alcpu.com/CoreTemp/?utm_source=CTGadget

Nick McDermott - 3dGM Admin

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

I just had my system give me the dreaded HTSF error again. I am going to try what you said.

VID says 1.3500v but I can't find the actual voltage it's getting.
The temps for the CPU cores is running hot on idle. I'm getting 50+C. Core 0 is around 50 and core 1 is around 58.

I did some more research about and I found a thread that had the HTSF error as the topic. One of the posts had an answer as to what was causing it. He said the motherboard isn't getting enough voltage to the CPU.

UPDATE: I just found a better thread that explains what is causing the issue. http://www.amdzone.com/phpbb3/viewtopic.php?f=38&t=137699#p183035

This guy seems to have the real answer: He says the hypertransport clock speed should always be lower than the clock speed of the cpu-nb. And if they are equal than the HT will sometimes exceed the CPU-NB clock and this caused the HTSF error.

EDIT: Well my HT is clocked at 1000 mhz and my CPU NB is 2900...so that's not the case. BTW I looked at my event log in windows and I am looking at the critical errors part and it shows all the crashes and shut downs that occured. I know for a fact that the ones on the 19th of Feb were due to power outages and the latest ones were HTSF errors. Around mid to late January was when I installed a new GPU and that is when I had major issues with it. It did not want to work with whatever Network card I put in as it would just freeze my PC. You can look at the thread which I made back then. I updated the BIOS on 1/25/2011 to the latest one (this could be the cause possibly). Another thing that happened around that time was I nearly fried everything in the computer by changing the voltage switch on the PSU on accident. So many variables. I installed the new PSU around Feb 13th.

eventcrit.png
dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

Great news. I put in a new CPU and the voltage fluctuation has stopped. Almost everything is flat lined with very minimal fluctuation.

dustyschaffner
dustyschaffner's picture
Offline
Joined: 06/02/2009
Posts: 334

I recently built a computer for a friend using the MSI 790XT also. The problem you are having is, in fact, from the motherboard. With my friend's computer, we managed to discover that the northbridge was overheating, causing it to just dump info and stop working "core flood error". We put a fan blowing directly onto the chipsets and the problem stopped instantly. Try this before you go spending more money.

eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

I've been seeing more and more issues with MSI and Asus boards recently. This is disturbing that these GREAT BRANDS seem to be losing their quality control.

My big rig is running an MSI board, but when I upgrade to an X6, I'm going Gigabyte, and it's purely because Gigabyte is still rocking the quality idea.

Nick McDermott - 3dGM Admin

dustyschaffner
dustyschaffner's picture
Offline
Joined: 06/02/2009
Posts: 334

It is a shame. I used to have good faith in MSI.
I have tried a Foxconn motherboard before. The off-brand name had me a little uneasy but it was a solid motherboard and a VERY good performer. Thing was quick as lightning and didn't have a single issue.

eire1274
eire1274's picture
Offline
Joined: 09/12/2003
Posts: 1221

Considering that Foxconn is a Chinese brand under ownership of Hon Hai Precision Industry Co Ltd., they have a step up in development due to higher input revenue. They used to make parts for the earlier Atari platforms, they now make the bulk of Apple products including PC mainboards, iPods, iPhones, and laptop chassis, as well as the entire XBox series and the Nintendo Wii, and the bulk of Nokia and Sony/Ericsson phones. I believe that even the latest generation of the PS3 is a Foxconn product.

They are a happy to make products that will be branded another name.

Nick McDermott - 3dGM Admin

dustyschaffner
dustyschaffner's picture
Offline
Joined: 06/02/2009
Posts: 334

I could have sworn Sony was Japanese..

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

I got that evil error again yesterday...alright I guess I have to cool the northbridge somehow. Where is it and how do I place the fan so it will blow cool air on it?

dustyschaffner
dustyschaffner's picture
Offline
Joined: 06/02/2009
Posts: 334

You should see a heastink near the middle of the motherboard, directly above the top PCIe slot. My friend and me managed it with just the case fans but he was using an HAF-932 case. I would suggest using an Antec SpotCool fan but it seems they are discontinued. Try your best with a case fan.

dhlord64
dhlord64's picture
Offline
Joined: 08/19/2008
Posts: 48

What was most strange about the error is that it happened when I was watching youtube. It didn't happen during intense gaming for long periods yet it happened while watching a youtube video when I just turned the thing on.