Having massive issues 8Pack games crashing out freezing/soundloops

Soldato
Joined
26 Jan 2004
Posts
6,278
Location
Scotland
I dont know where to turn or what to do hoping you have some idea.

I have 2 his cards 7950s
i7 @4.5 4770k
16gb 2133 cl9 avix
creative ZXR
---------------------------

i can cant seem to play games for long.
game freezes, sometimes with recovered ect and most of the time i can get back to the desktop to look into the eventviewer.
and it says failing module aticf32.dll or something like that meaning graphics card driver.
im using the latest drivers i have tested both cards single and there is no issues it only happens in crossfire.
------------------------------

i can not play BF4 or Splintercell blacklist for more than 10-15 minutes before crash freeze or crash freeze sound loop.
i was just playing tonight a none demanding game cs go and gpus where barly getting used
and i got a freeze/soundloop.

-------------------------------

i have 3 possibilitys
the cards need RMA`d to be replaced or the motherboad need replaced or the memory needs replaced, i had these freezes and issues even before i used your OC settings....
so unsure if its the memory.

got any settings i can try cause im stuck......
also sometimes when im booing windows right after the login screen i will get a BSOD116 saying that graphics cards failed to recover :/
and when i do get in i get the above issues.
 
116 is defo grapohics. Remove all overclocks. Try PCH at 1.2v. Disable HPET. Try again. are you trying to overclock the cards in Xfire?? If so disable uLPS.

PCH 1.2v is that the IO or the Core ? just to make sure
and this hpet setting what is it and do people disable it what benifit does it have and doesnt have does it break performance ? or anything...
i already have UPLS disabled.... tested the cards in oc mode and stock mode same issue so cant be the overclock on the cards. as my oc is mild 1070/1400 from 1120/1450...
memory on my cards do not oc as there using Hypida memory or something like that.

EDIT : ok mate i have changed all but the hpet setting i cant seem to locate it in my bios, is there even a setting HPET in my bios ? gigabyte OC board ? i will let you know if the issue is cleared up...
also changed the CORE PCH and left the IO @ default auto hope thats correct.
 
Last edited:
It could also be your PSU. If the cards are fine singular, then there is obviously less load on the PSU than with xfire. Do you have a spare you can try?

I dont have a spare mate :(
the PSU is not even 4 months old. brand new direct from pc power and cooling mkII 950w psu 1 single rail.

I dont think its the psu as i did not have issues like this with my x58 system and all thats been changed in the past month is the CPU/MOBO/RAM thats it and the x58 system had no issues at all, and it also uses more power draw compaired to haswell z87 chipset.
 
Update, seems teh PCH Core @1.2v fixed it, played splinter cell for 1 hour 20 minutes and zero freezes or crashes @1120/1500 Xfire.
---------------------------------------------
Its weird how the default voltage on the motherboard has issues with my cards though and requires massive amounts of PCH voltage to stabilise it and i also plugged in the 6 pin to give extra juice to the pci express lanes on the board... very strange....

im still intrested what this Event timer does and where that setting is in my bios cause i cant locate it.

cheers again 8pack just change PCH seems to have fixed the issue :o though i will report back if there is a hicup....
 
UPDATE :(

well so much for that, i thought everything was ok, for the past day and half its been fine then bam, same issues reappearing :( the very odd 116bsod on login happens rarely now not as much as it did when i changed PCH 1.2v it cleared it up but its still there much rare though

but still my system is freezing up in BF4 and SPLINTER.
i even tested splintercell in windowed mode and it still freezes up

what happens is the game will freeze i will see the gpu jump from 50% to 99% then boot me back to desktop with display has been recovered most of the time i have to hard reboot.

i personally thing the GPUS are broken or one of them is broken or the motherboard
i cant afford at the minute to send all this stuff back to OCUK then for OCUK to say no issues and send them back arghhh so annoying not to be able to play a game properly.
 
Try each card on ts own and also remove all overclocks on the system while testing.

I would also force gen2 in bios and disable HPET.

ok i will give those a try,
am i loosing a lot of performance going 2nd gen compaired to 3rd gen pciex ?
------------
also i run a memtest on the Overclock settings you gave me
i did 2 loops and no errors....

but when i test, test 6 for more than 2 loops it starts to show up errors :o
can memory instability cause the issues im having with graphics ?
 
Nope!! Why would I tell you to do it if you was?? Look at me world records in vantage etc most are at Gen 2!!!

hehe ;) ok gen2 i will enable, also about the memory ? is there a setting i can relax cause test 6 is spitting out errors after 3-4 loops even though the first 2 loops pass without issues.
just wondering if its a voltage thing ?

also with OC boards the digital io and analog io are in voltages of 0.00+v
and i have them both set at +0.003v is that to high ? or should i shoot for higher ? though not sure what the safe max or average is for these voltages.

thanks again for your help mate.
 
Post up the memory settings man!! I cant tell with nothing to work with. IOA and IOD can be +0.2v no problem so your hardly killing the chip ;)

HI mate i posted the settings i use for memory in this thread : http://forums.overclockers.co.uk/showthread.php?t=18544792

there the settings you tweaked.......

--------------------------------------------------------------

here is a image i took of my voltages :
39pc.png



sorry about the bad pic must have had something on the lens lol
the top voltage thats hard to read is 1.860v for vrin.

--------------------------------------

running 4.5 with 4.3 uncore my crashes and graphics issues have completely gone
when i revert my memory back to XMP profile no errors in memtest no issues.
but there must be a setting im thinking voltage that is causing the memory to fail after more than 2 loops of test 6....
i can complete a full test no errors but if i set the option to run test 6 over and over and over it will error out after 3+ loops.

any guess`s or is my memory just not as good as the stuff you tested.
 
Your SA, IOA amd IOD are still incorrect + 0.2 not 0.02!!! Come on man!!!

ok mate tell me if this is correct now cause im still getting issues :(
the first voltages i get errors in test 6
the 2nd voltages i get errors poping up in test 6 after 4 complete no errors passes then errors popup :/

---------------------------------------------------

nq4c.png



no idea what is causing this could be i just do not have good stuff or my motherboard is not as good... or the IMC on my 4770k is crap.
2133 speeds are fine though, if we cant sort this is there any chance you could do me a tweaked settings for 2200 aswell as 2133 please, as i dont think im going to get 2400.
 
Last edited:
My sticks passed several loops at the settings I gave you for 2400mhz. It is possible your sticks are not as good.

Try 1.7 Vdimm. Hynix IC will be fine 24/7 at this.

OK mate i will report back, the voltage on the ram on the above was 1.65 as 1.68v in windows
will try 1.70v and report back. and yeah mate my ram passes several loops but when testing out test 6 only in a loop after 3-4 error free passes then errors appear as it shows in the above picture its very strange as the memory should not pass 1 or 2 without spitting errors but it does pass and then spits :/
 
update: hi mate still broken :( set 1.690v = 1.704v in windows
run memtest, granted it goes much further this time with no errors, 9 complete pass`s no errors then errors popup :/

--------------------------------------------------
b09o.png


could just be that to get higher than 2133 or 2200 i need better memory, i shoulnt have to pump a lot of SA/AI/DI voltage for 2x sticks of 8gb and even with that its not stable and 1.7v still does not make it stable i dont know what else to try.....

might have to get ya to do a 2133/2200 profile for me to try... if im not bothering to that much that is lol i feel as though your getting fed up with me hahah
just a thought im pushing 1.250v into vring to get 43uncore stable, could it be 1.250v is not enough for 43uncore and thats what could be causing the errors, though im just guessing here.
 
Last edited:
update to the previous post, this is very weird this one lol let me know what you make of it.
-------------

tests where run at default 1.65vram voltage
and the voltage settings in the picture.
also updated to the final version 5.1 memtest86.
--------------

the first run was done without multithreading... loads of passes no errors at all tends to make me beleave there is nothing wrong with your OC tweak settings or voltages.
but then the 2nd run was with SMP Multithreading enabled.
loads of errors in test 6 yet again :/

makes i thought it was cause of vcore not enough juice so i bumped it and still errors
then i tryed vring and bumped that and yet again errors.

what could be causing this ????? could i have a flaking broken cpu ? when multi is enabled ? with ram ect.

z49b.png


as you can see in single thread mode no issues no errors you would think if there was no errors in single mode that the memory would not give errors when in multicore mode also you can see in single mode the memory runs faster than in multi mode, makes no sence.
 
I cant remember exactly how your timings went. Maybe you can relax TWCL or TRRDRD.

Hi mate, here are your settings i used : http://forums.overclockers.co.uk/showthread.php?t=18544792

I cant find in there TWCL or TRRDRD they must be named slightly different or i could be blind as im tired lol

EDIT : found TWCL its called TWL on my board, now i need to find that TRRDRD lol

EDIT2: TRRDRD is found now its called TRDRD_DR

are those the correct ones to bump up a notch 8 pack ?
 
Last edited:
TWL 8
TRDRD 5 or 6

Hi mate,
tested out by changing just TWL from 7 to 8 and i could not get the pc to boot
when i changed 1 stick to 8 and kept the other stick at 7 then the pc would boot

also tested just changing TRDRD_DR from 5 to 6 and pc booted but still errors in test 6.

im not sure but i think one stick is weak in this set for some reason... and that could be the reason why i can not get your settings working without errors.

only 2133 works. or 2200 works. but i need a set of tweaked settings for those speeds.
if possible cause i dont think im going to get stable at 2400 :( might have to buy a high end kit as i will have a higher chance of getting higher with a higher set, thanks for the help so far though... much appreciated.
 
Hi 8pack mate, i tried your method and changed TRFC from 224 to

240 = errors test 6
250 = errors test 6
then i relaxed the setting to its default 279 or something like that and there is still errors
so i have no idea what to do next.

is there any chance you can test a set that i have and i will put a order in for another set then i can send these ones back for a refund or is that not possible.
 
If they work at default then you cant get a refund no overclock is guaranteed and it could be IMC, board and many other factors. Set defaults and mem test.

ok, though at default the memory works fine also at XMP the memory works fine.
do you have a tweaked settings i can try for tighter 2133 speeds or 2200 speeds.

if possible. cheers.
 
Back
Top Bottom