• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

PC GAF: I could Really use some help. A free Steam game goes to the GAFFER with a fix

K

kamspy

Member
I don't know where to start, so I'll post some facts first, and then some things I've already tried.

I'll be more than happy to shower people in Steam gifts if they can help me out.

The theme of this thread is : Hyper Flood Sync Error


Facts

-My PC crashes whenever I load a graphically intensive game. Crysis, GTA4, L4D2 (?)

-My PC doesn't crash when I load mid level games. Games that work include FEAR, Lara Croft GoL, Mirror's Edge.

-I can run Furmark and Intel Burn Test with no crashing.

-When it crashes I just get a black screen and the computer reboots. A prompt at the POST screen reads HYPER FLOOD SYNC ERROR, PRESS F1

-I leave my PC running 24/7 usually and woke up to these errors probably twice in the last year. Didn't think much of them.

-Dec. 30th I had a marathon gaming session going. Probably 4 hours straight of GTA4, and then I launched Borderlands which gave me my first gaming related HYPER SYNC FLOOD ERROR. I was unable to boot to Windows afterward. Safe mode worked, but only a reinstall got me back in to regular windows.

-I was undervolting my CPU for the last three weeks or so to combat high temps. It was running at 1.275v. Stock is 1.4v. Ran stable before this error. Not sure if it's related, but google gives me a lot of voltage related stuff when searching.



SPECS

ECS A780 GM-A v1.1 Black Edition
AMD Phenom II 955 BE Black Edition
BFG Nvidia 275 GTX OC (factory oc)
4 GB G.SKILL DDR2 1066 2v
BD/DVD-R Pioneer drive
500GB WD HDD (Boot)
1 TB Seagate HDD (Media)
750w Corsair PSU

Windows 7 Enterprise x64


This build ran stable for a very long time. Never even gave me a BSOD before this. Everything is less than 2 years old aside from the motherboard.



Things I've Tried

-Reinstalling windows on a freshly formatted drive.
-Updating BIOS
-Using default BIOS values
-Using default RAM voltage and clocks in BIOS (recommended voltage and clocks are different)
-Boosted Voltage and under voltage CPU, and NB by 0.25. Just about every voltage combo I could find. Seems to be the common theme on google.
-6 passes of memtest.
-Intel Burn Test and Furmark run fine.
-Downclocking my factory overclocked GPU to "stock clocks" for that model.



So I'm stumped. This is only happening with games. I even tried some full screen Blu Ray playback and it worked fine.

It seems like the Motherboard is an easy target. Cheap maker, pretty old, thousands of hours of use. The thing that's got me rethinking ordering a new mobo is the fact that almost every single HYPER SYNC whatever error is related to a Phenom II 955 or 965 (which come off the same wafer. Same exact chip with different firmware basically).


I don't have a lot of money GAF. I just spent a ton during the Steam sale and on other holiday gaming expense.

I even resorted to playing RalliSport challenge on my Xbox 360 just to get my 60fps racing fix. I need my gaming PC. I need my Crysis. I need my DIRT 2.


I'm in dire need. I'll post any DXdiags, whatever. You recommend it, I'll try it. Info you need, I'll post it.
 
P

-Pyromaniac-

Member
Looks like a mobo problem. Looks like it can't handle your 1066 memory. Try putting in 800s instead. Or try putting in only 2 gigs of what you have if possible. I think it's a problem around that, just experiment.
 
P

-Pyromaniac-

Member
Never trust an asterisk =p

edit: and I mean he has tried so much and there is not a whole ton it could be, so it's worth a try I'd say. At least it'll narrow things down.
 
B

beje

Banned
The first things you must do on suspected overheats are...

a) Clean your computer insides! Get a compressed air tool specifically for this task. If you don't have one of those available, combine your vacuum cleaner (just the tube, not the end accessory you usually use for the floor) with a hair drier (in "cold" mode, please) to move the dirt in every direction possible so it's easier to capture. Check EVERY single place, especially between coolers' fins. One time, I had to literally pull curtains of dust and dirt with a pair of tweezers between the CPU sink fins when I was cleaning a friend's PC. An old brush will do wonders for the finishing touch.

b) Remove the CPU sink so you can clean the old thermal paste and apply new paste. If you have never done this, or you suspect you may be doing it wrong, youtube is your friend.

c) Do the same for your graphics card's GPU. This is much more delicate and will be completely different deppending on the model, so operate with extreme care and be sure to memorize the steps you take so you can build everything back with the correct screws. Take into account this may void your warranty!!!

Anyway, for everybody else, you should be doing the (a) part every three months and (b) and (c) should be considered as annual or semestral maintenance (depending on the weather, temperature of the room, etc...) even if you don't have any kind of problem.
 
W

Woakes

Member
Clean out the graphics card, make sure it's free of dust. Blow it out with canned air, or at a push use a straw and blow (don't spit). Hot Nvidia's tend to black screen of death without any apparent visual artifacts before it crashes. Just a thing to rule out.
 
K

kamspy

Member
I've tried running the RAM at 800Mhz.

It's 800Mhz RAM anyways. G.SKILL are cheap bastards who sell RAM that needs to be overvolted to reach the advertised speeds.

Keep in mind this computer has ran thousands of gaming hours perfectly stable for a couple years now with mostly the same parts. Newest thing is the 275 I bought last Feb.


I'm so OCD. I have a big backlog of older games that run. I probably wouldn't even play Crysis or Metro. Just know I can't is what's killing me. It's been the focus of all my free time since it happened.

Woakes said:
Clean out the graphics card, make sure it's free of dust. Blow it out with canned air, or at a push use a straw and blow (don't spit). Hot Nvidia's tend to black screen of death without any apparent visual artifacts before it crashes. Just a thing to rule out.
Click to expand...


Hmmm. I did give it a spray when I was taking out parts to rule out a stick of RAM of the Nvidia. Problem is my integrated graphics aren't even capable of starting the games the Nvidia can't run. I don't think I can troubleshoot with it.

Any ideas on how I could isolate the card? I wonder if BFG can even replace one these days. :( So many new games.:(


FWIW, I set my fan profiles to never let the card exceed 85c for as long as I've had it. Metro might have tinged 90c a few times, but that was months ago.
 
V

vocab

Member
What's the temps look like in Crysis, and GTAIV?


Try turning off cool and quiet. I heard some people fixed HYPER SYNC FLOOD ERROR by doing that. AMD/ATI love their power saving options that always have problems.

Since you crash during CPU bound games, I would think that could be a conflicting problem.
 
M

MattPeters

Member
When you changed the voltage, what did you go up too?

Edit: nvm, just reread your post and found the volt stuff. Derp I'm a tard
 
B

beje

Banned
kamspy said:
Any ideas on how I could isolate the card? I wonder if BFG can even replace one these days. :( So many new games.:(
Click to expand...

If you still haven't done the general cleaning: do it, trust me, your computer will be grateful. To isolate the graphics card, this is as easy as swapping it with a different one from a friend. If this is not possible, maybe you can get some help in the shop you bought it if you ask them to lend you a spare one they have around for their stuff to do some cross-test.
 
S

Seagoon

Member
K

kamspy

Member
The case stays dust free. I just repasted the proc about a month ago when I got rising CPU temps. I've monitored CPU voltage while gaming (before this all started) and it's just the standard droop.


I don't have a spare card laying around. The onboard is a ATI HD3200. I don't know if it's even enough to launch the game that are crashing my comp right now.

My CPU temps never break 60c and GPU temps never break 80c-83c. I run SpeedFan and Afterburner almost every time I play a game just to see what my temps, vRam usage and stuff like that is. I'm a nerd about that stuff.

EDIT:

Memory voltage is at 2.1v right now. I've tried the BIOS default 1.8v too. Neither stop the crashes. I got to the EA logo in Warhead once in about a doezen attempts. Never make it past a black screen in other games. Just Cause 2 just gives me a black screen but never crashes the comp.

3 days ago this computer ran flawlessly. I haven't made any changes or installed anything other than games since then. Aside from the Windows reinstall mentioned in the OP.
 
P

-Pyromaniac-

Member
Left 4 Dead 2 will be able to run with the ati hd3200. Might be slow or look like shit but I know for a fact the demo can run on that. So the game should launch. If that works then it's vid card related.
 
K

kamspy

Member
Alright, I'm gonna uninstall the Nvidia drivers and pull the card. See if the onboard gives me any luck.

I'm actually hoping it crashes. My 275 has been nothing but good to me. :|
 
D

DodgyGeezerFella

Member
At a glance i would say motherboard. A friend of mine has been having crashing on intensive games as well, but been playing mid-range games for hours no problem. FurMark & Intel Burn Test gave no errors too. Replacing everything in his system bar the motherboard, freshly formatted, the problem still exists, we plan to change the motherboard soon, its the only thing we can think of thats causing the problem.

Maybe its a similar issue with yours, just throwing my 2 cents in there. If possible, try all the components on a different motherboard and see how it goes.

I feel your pain :(
 
DonMigs85

DonMigs85

Member
-Pyromaniac- said:
Well at least you'll know. If it still crashes then I'm gonna go with memory or PSU problem.
Click to expand...
Methinks if uninstalling graphics drivers or switching to the onboard doesn't help then it's likely either the mobo, RAM or CPU (I think the Hyper Transport controller is embedded into it).
 
E

Easy_D

never left the stone age
Ogs said:
At a glance i would say motherboard. A friend of mine has been having crashing on intensive games as well, but been playing mid-range games for hours no problem. FurMark & Intel Burn Test gave no errors too. Replacing everything in his system bar the motherboard, freshly formatted, the problem still exists, we plan to change the motherboard soon, its the only thing we can think of thats causing the problem.

Maybe its a similar issue with yours, just throwing my 2 cents in there. If possible, try all the components on a different motherboard and see how it goes.

I feel your pain :(
Click to expand...
My brother had the same thing but with World of Warcraft. Replaced his GPU with my 8800GT when I got my new GPU. Shit worked just fine.

A few months later the GPU was broken, overheating had damaged it to the point that the moment he started the game, there were artifacts of hell, then the computer would crash.

So I had to put his old GPU back in, strangely enough, the game worked just fine for a while.
Then it started happening again. Strange shit indeed. Then I noticed there was a newish BIOS update that fixed random crashes, flash his shit, now it works just fine again.

My point is, computers are fucking strange and no computer behaves the same as any other. It's a high price to pay for sure but god damn is it worth if when everything works as it should.

Sorry I can't be of any help, OP :(
 
P

-Pyromaniac-

Member
Yeah the mobo is definitely the easy suspect but everything else works flawlessly and it just happens to be the intensive games that cause issues? That's the weird part that makes me think otherwise.
 
DonMigs85

DonMigs85

Member
-Pyromaniac- said:
Yeah the mobo is definitely the easy suspect but everything else works flawlessly and it just happens to be the intensive games that cause issues? That's the weird part that makes me think otherwise.
Click to expand...
Maybe only the intensive games fully saturate the RAM bandwidth or the Hyper Transport link.
 
E

Easy_D

never left the stone age
-Pyromaniac- said:
Yeah the mobo is definitely the easy suspect but everything else works flawlessly and it just happens to be the intensive games that cause issues? That's the weird part that makes me think otherwise.
Click to expand...
Googling gives me some probable causes for the Hyper Transport Sync Flood stuff.

Apparently setting the Hyper Transport Link clock higher than the NB Frequency (Someone with knowledge might want to explain what this means, I have no idea :lol )

South Bridge Overheating

Or.

Your memory not getting enough juice - Makes sense when you consider that more demanding games probably use up more memory.


Now I just found these and translated it to english, I don't know anything, but if it helps, it helps.

Edit: Considering it ran stable for a good while it's more likely something is just plain old broken though :( then again PSU's losing their JUICE MAKING QUALITIES over time might factor into it.
 
S

Seagoon

Member
Aside from it being the motherboard - I can't help but wonder if it is the PSU at fault. I've worked on machines with power supplies that work fine for the most part but expose faults when under load.
 
L

legend166

Member
I had this problem years ago with my PC. It turned out to be a dodgy power supply. We got it swapped around and it fixed it.
 
alf717

alf717

Member
This usually isn't an easy issue to resolve since it could be one of or even many things.

What about the temps on your machine? Something could be overheating which could be the GPU, CPU or PSU.

Check your board for swollen or popped capacitor which could cause random shut offs.

Run memtest+ and see if you have any read errors. One read error means your RAM has an issue. Although memory issues usually result in blue screens or the machine not even POST'ing I would run this test at least once

Do you have a multimeter? Might want to test the PSU and see if it has any defects. I have this issue when an ECS board as well. Graphics intensive games cause it to crap out. No blue screen what so ever it just shuts off. I believe it is my CPU since it idles at 36+C's and that is with cool n' quiet turned on.

Edit:

Sounds like the CPU found this on another forum

EDIT: OKAY, FOUND OUT THE CULPRIT of this Hypertransport Sync Flood Error.

Did some testing on the CPU, CPU-NB, HT and Ram.

The CPU was ONLY component that reproduced the HTSF Error.

When not enough volts are being provided to the CPU it will reboot at any time (Especially during heavy to max loads) with either a black screen or verticle lines right before rebooting. While posting you see this messege.

"hypertransport sync flood error has occured. Press F1 to resume"

So there are a number of Scenarios that may cause this BUT they ALL lead back to the main reason. Not enough power to the CPU at the time that the heavy load produces failure.

1. To high of an OC without the ample amount of CPU volts.
2. Plain, not enough CPU Volts at any speed.
3. Possible Bad power supply.
4. Motherboard with very high amount of VDroop which ultimately leads back to "Not Enough CPU Volts". This will reveal itself with a load on the CPU. Maybe not during typical use. Could just be that the motherboard has CPU power providing issues as well (Good power from PSU but motherboard wastes the power before reaching the CPU).


Hope that this helps someone out!
Click to expand...
 
K

kamspy

Member
No dice on the integrated video. Still get Hyper Transport Flood sync errors launching games. :/

I did stress the GPU the best I could with Mirror's Edge. PhysX and 8xMSAA. No artifacts, nothing. Smooth as ever. Try to launch Black Ops, Crysis etc. I get the crash and the error.

Anyone do googling and notice that it's always the 955 or 965? That's the one thing stopping me from just ordering a new mobo.

Microcenter does a $20 CPU diagnostic. Seems like it might be a waste of money. Anyone have any experience with it?
 
alf717

alf717

Member
I'm willing to bet it could be the 4 pin ATX connector, most higher end CPUs, like X4's and X6's, require a board that uses all 8 pins unless it is less than a 125watt CPU. You might be able to just replace the board with a higher end version and be good to go with the rest of the setup. Only issue is unless you go second hand a new board will require DDR3 RAM unless you can get a nice deal on an AM2+ board.
 
W

Wallach

Member
Jin34 said:
I may have found your problem. I googled your board and found this in the product support page: Note: This board supports CPU up to 95W TDP only; you can refer to AMD website to check your CPU.

And your CPU is a 125W CPU.

http://www.ecs.com.tw/ECSWebSite/Product/Product_Detail.aspx?DetailID=892&CategoryID=1&MenuID=20&LanID=0
Click to expand...

I was actually just thinking about this. Have you looked at the area surrounding the socket? Look to see if there's any sign of singeing or carbon build-up on the MOSFETs or something. I used this board for some time with an X2 6000+ and didn't have any problems but I always worried that the quality of the MOSFETs was going to bite me in the ass down the line.
 
K

kamspy

Member
Jin34 said:
I may have found your problem. I googled your board and found this in the product support page: Note: This board supports CPU up to 95W TDP only; you can refer to AMD website to check your CPU.

And your CPU is a 125W CPU.

http://www.ecs.com.tw/ECSWebSite/Product/Product_Detail.aspx?DetailID=892&CategoryID=1&MenuID=20&LanID=0
Click to expand...

Yeah, I know. But on the old ECS website when I bought the 955 it listed the 955 and 965 as supported chips. Probably just to help sell the board really.

FWIW, I used a 125w CPU in that board before my 955, (a 6000+). Been running that board since it came out.

Since it's technically out of spec, it's hard to troubleshoot. Probably gonna go look for a AM2+ board tomorrow. I see no reason to upgrade to DDR3. Gonna try and ride the AM2+ platform til the next revision in desktop memory. That was my plan anyway. I hope I can find a board. Newegg gives me like 5 options and 3 are microATX. I'm scared my 275 will warp such a little board. It may be an old card, but it's huge.


Also, how would I properly dust that GPU? It's entirely enclosed, and IIRC, taking off the enclosure voids my warranty (EVGA from now on).

I really appreciate all the help guys. Thank you.

Wallach said:
I was actually just thinking about this. Have you looked at the area surrounding the socket? Look to see if there's any sign of singeing or carbon build-up on the MOSFETs or something. I used this board for some time with an X2 6000+ and didn't have any problems but I always worried that the quality of the MOSFETs was going to bite me in the ass down the line.
Click to expand...


I'm a EE with a focus on pcb devices and the part quality is what made me buy this board. All the caps etc. are really nice for the price. The layout is really good too. I think I might have warped it after all the years though.
 
alf717

alf717

Member
kamspy said:
Yeah, I know. But on the old ECS website when I bought the 955 it listed the 955 and 965 as supported chips. Probably just to help sell the board really.

FWIW, I used a 125w CPU in that board before my 955, (a 6000+). Been running that board since it came out.

Since it's technically out of spec, it's hard to troubleshoot. Probably gonna go look for a AM2+ board tomorrow. I see no reason to upgrade to DDR3. Gonna try and ride the AM2+ platform til the next revision in desktop memory. That was my plan anyway. I hope I can find a board. Newegg gives me like 5 options and 3 are microATX. I'm scared my 275 will warp such a little board. It may be an old card, but it's huge.


Also, how would I properly dust that GPU? It's entirely enclosed, and IIRC, taking off the enclosure voids my warranty (EVGA from now on).

I really appreciate all the help guys. Thank you.




I'm a EE with a focus on pcb devices and the part quality is what made me buy this board. All the caps etc. are really nice for the price. The layout is really good too. I think I might have warped it after all the years though.
Click to expand...

But the FAQ does state

Question
Can A780GM-A V1.1 support 125W CPU?

This model is designed for supporting 95W CPU base on AM2+ platform. As you can see there are some 125W CPU listed on CPU support table,
these CPU are validated by ECS and able to work properly. Since it is over specification, so there is an asterisk (*) in the table to identify and remind user friendly.


* means validated by ECS at time of writing. Changes to component or CPU may produce a different outcome. The result shown is for reference only and is subject to change w/o notification.

Sorry to say it but you are going to need a new board. It could be that the stepping/revision used were different on the chip they tested versus the one you have. I wouldn't fumble around with this issue anymore and just get yourself a newer board.
 
W

Wallach

Member
kamspy said:
I'm a EE with a focus on pcb devices and the part quality is what made me buy this board. All the caps etc. are really nice for the price. The layout is really good too. I think I might have warped it after all the years though.
Click to expand...

I think most of that stuff is high quality, but it's clear the board was designed mostly in the 95w TDP range. I don't know if the MOSFETs (which are bare on that board if I remember right) were really designed to run 125w CPUs for such a long time.
 
J

Jin34

Member
Ram prices are real low again just in case you thought they were still really high and that was your aversion to an AM3 board.

I get totally ramdom crashes on mine but I did have that one incident where I bought a toslink module thinking all the plugs on mobos where standard and it turned out the ground/voltage pins were backwards and lo and behold the plastic connector cover fused into the module and some nice burnt electronic smell as well as a short when the module tapped the case :lol

It turned out fine afterwards but many months later the crashes started filled with cmos checksum bad errors.
 
alf717

alf717

Member
kamspy said:
I just wonder why board ran for so long so well. It was stable with 125w CPUs for a couple years.
Click to expand...

The only other option is to test the PSU you can jump start the PSU and test it with a multimeter or buy a device to test the voltage of each connector. If you have a multimeter you could at least test the PSU and make sure it isn't the issue but the fact that the board only supports 95 watt CPUs is most likely the cause. If you are interesting in testing the PSU I'll see if I can find you a guide just make sure you don't shock yourself.

Edit:

Here you go, have at it if you want:

http://www.overclock.net/faqs/96712-how-jump-start-power-supply-psu.html

Again, use caution that you don't shock yourself.
 
K

kamspy

Member
alf717 said:
The only other option is to test the PSU you can jump start the PSU and test it with a multimeter or buy a device to test the voltage of each connector. If you have a multimeter you could at least test the PSU and make sure it isn't the issue but the fact that the board only supports 95 watt CPUs is most likely the cause. If you are interesting in testing the PSU I'll see if I can find you a guide just make sure you don't shock yourself.
Click to expand...

I'll give it a try. Already found a nice AM2+ mobo for $65. It's mATX.

Anyone know of any problems with mATX mobos and big, heavy hot GPUs?
 
W

Wallach

Member
kamspy said:
Everyone who helped shoot a Steam invite to kamspy.
Click to expand...

Appreciate the sentiment but I'd rather you put that money towards your new motherboard!

When you get to taking it out eventually, check the MOSFETs as I'm curious if they show any significant signs of wear after using that CPU for so long (you probably can't see them if you have a decent sized HSF on there right now).
 
J

Jin34

Member
kamspy said:
I'll give it a try. Already found a nice AM2+ mobo for $65. It's mATX.

Anyone know of any problems with mATX mobos and big, heavy hot GPUs?
Click to expand...

Only problems I know is that mATX boards tend to have lower CPU wattate than full ATX boards so make sure it supports a 125W cpu.
 
alf717

alf717

Member
Wallach said:
Appreciate the sentiment but I'd rather you put that money towards your new motherboard!

When you get to taking it out eventually, check the MOSFETs as I'm curious if they show any significant signs of wear after using that CPU for so long (you probably can't see them if you have a decent sized HSF on there right now).
Click to expand...

Ditto, save your pennies and let us know the results of PSU test and status of the board. I've got a swollen capacitor on my board which seems to be causing issues, also have a bad IDE port. I'm hoping 2011 is the year I find work cause this PC is falling apart on me. :lol
 
K

kamspy

Member
Preciate it fellas. I'll be going to microcenter tomorrow I'd love to find a AM2+ board. I will update the thread with results and pics of hardware porn.

This little ECS has lasted me a grip though. For $59 3 years of good service is OK in my book. Especially considering I ran it out of spec the whole time.


I've heard bad things about boards with Nvidia chipsets. Would I be right to avoid them?
 
M

markot

Banned
Nvidia + AMD seems like a bad combo :p

Do you need a microATX though? Id just go for a normal one.

But yea, I remember when my computer was screwing up on high power/resource games, and my friends, the problem was the PSU in both cases.
 
You must log in or register to reply here.

Similar threads

13900K and 14900K Instability Causing Oodle Decompression Failures
Hardware 2
80
7K
winjer replied
[Verge] Windows on Arm isn’t ready for gaming
22
2K
LordOfChaos replied
  • Poll
New PC build. Rate, opinions and help wanted.
45
3K
Windom Earle replied
PC Gaf, I need some help with my PC
28
1K
Kurotri replied
[Jez Corden] If Microsoft really IS bringing Steam, Epic Game Store, etc. to Xbox consoles, it needs to answer these big questions
2 3 4
155
13K
sachos replied
Top Bottom