Oh Dear, more ATI 58xx mysteries!!

[LegendaryDread]

"The Radeon HD 5870 has 1600 stream cores which are all usable, while the Radeon HD 5850 has 1440 usable stream cores, because 160 out of the 1600 total stream cores are disabled in order to disable defective areas of the chip in a process called product binning so that defects which ruin some of the stream cores do not ruin the whole chip."

http://en.wikipedia.org/wiki/Evergreen_%28GPU_family%29

Could it be possible they're not checking that the chips are working in an effort to get more cards out faster?

Sorry if this makes no sense, I'm not really clued up on the whole GPU manufacturing process.

 

[Hairybudda]

While it's a possibility I'd say it's highly unlikely

 

[LegendaryDread]

Yeah, I didn't think it likely either, until a clanmate received his third DOA 5850 yesterday.

 

[kylew]

"The Radeon HD 5870 has 1600 stream cores which are all usable, while the Radeon HD 5850 has 1440 usable stream cores, because 160 out of the 1600 total stream cores are disabled in order to disable defective areas of the chip in a process called product binning so that defects which ruin some of the stream cores do not ruin the whole chip."

http://en.wikipedia.org/wiki/Evergreen_(GPU_family)

Could it be possible they're not checking that the chips are working in an effort to get more cards out faster?

Sorry if this makes no sense, I'm not really clued up on the whole GPU manufacturing process.

Binning is something that's used on most GPUs where they go into multiple graphics cards.

Any cards you see that have the same GPU but different amounts of shaders will have been speed binned. Unless it's not necessary, but most of the time, it is.

 

[hyperst]

Yeah, I didn't think it likely either, until a clanmate received his third DOA 5850 yesterday.

to receive three is very unlucky, but i would look more into the actual rest of the system at that stage as to receive three is highly unlikely, unless where there getting rma'ed, vendors are sending out the rma'ed ones to test with other peoples setups

 

[LegendaryDread]

That explains a lot Yeah, they were RMA'd

 

[kylew]

to receive three is very unlucky, but i would look more into the actual rest of the system at that stage as to receive three is highly unlikely, unless where there getting rma'ed, vendors are sending out the rma'ed ones to test with other peoples setups

Most definitely, 3 in a row, you know it's a case of PEBKAC.

 

[gareth170]

I have to agree - we can't absolutely rule out issues caused by flashing a different firmware - even after apparently returning to a stock configuration.

all of the 5850's are the same with the same bios/firmware, e.g a msi 5850 is the same as other brands with the same components and with the same bios. the msi 5850 unlocked bios just as higher ccc limits.

to the OP maybe 1 component on the card meant be faulty

 

[kylew]

all of the 5850's are the same with the same bios/firmware, e.g a msi 5850 is the same as other brands with the same components and with the same bios. the msi 5850 unlocked bios just as higher ccc limits.

to the OP maybe 1 component on the card meant be faulty

The simple act of flashing a BIOS can introduce problems if the flash chip sctors become damaged or something.

 

[gareth170]

The simple act of flashing a BIOS can introduce problems if the flash chip sctors become damaged or something.

but if all his issues was happening before he flashed it then it can't be because of the flashing.

 

[elpedro]

The simple act of flashing a BIOS can introduce problems if the flash chip sctors become damaged or something.

Surely the flash wouldn't be successful if that was the case?

 

[Hairybudda]

Would a bad flash not make the cards bios spit out checksum errors?

 

[hyperst]

well if you get a bad flash with a mother board its not allways apparent at first, sometimes just reflashing can sort it aswell, its allways a risky buisness flashing

 

[gareth170]

tbh the first question to the OP should be did he have issues before he flashed it

 

[hyperst]

Yes tried original bios.

your spot on, there is no consistency.

worked fine on win7 for weeks > then wont run on win7 > ran on vista > then wont run on vista > now works on win7 (dont know how long though)

well the worked fine on win7 for weeks would say it was ok when he first had it, dont know at what point it was flashed

and when i first heard talk of flashing bios for these new cards i did say i wonder who will be posting about borked cards

 

[gareth170]

i really don't think this issue is due to flashing. theres a lot of people that have flashed they cards without any issue...

hardware can slowly become faulty at anytime. not every card is made perfectly, maybe 1 component on his card meant be slowly on its way out.

 

[toon_mad]

i can assure you all this was not because of the flashing, i did say this already:

http://forums.overclockers.co.uk/showpost.php?p=15288690&postcount=110

http://forums.overclockers.co.uk/showpost.php?p=15294054&postcount=191

 

[gareth170]

i can assure you all this was not because of the flashing, i did say this already:

http://forums.overclockers.co.uk/showpost.php?p=15288690&postcount=110

http://forums.overclockers.co.uk/showpost.php?p=15294054&postcount=191

yeah thats what i been saying all along its not because of the flashing..

the thing i was think it could be is maybe 1 component on the card meant be slowly on its way out. just because it's new doesn't mean it can't have weak component(s).

 

[toon_mad]

well its still working ok for now in windows 7, im dreading switching the thing off and turning it back on incase the drivers dissapear and will no longer install again.

it is defo not the flashing, as said before, people were getting cards that would not boot at all, and some had fans revving up to 100% and then like myself, have a good card - when it wants to be good, others have got very lucky and had no faults at all.

 

[QwaF]

Not had a single problem with my two 5850's on Win7