4870 Overclocking Thread

[dnottis]

My first HD4870 was an HIS. It would crash before 1 min in Furmark. After 2 weeks it died. I believe it was a sign of a bad card.

[SNiiPE_DoGG]

My first HD4870 was an HIS. It would crash before 1 min in Furmark. After 2 weeks it died. I believe it was a sign of a bad card.

did you remount the heatsink, adjust fan speeds?

[dejanh]

My first HD4870 was an HIS. It would crash before 1 min in Furmark. After 2 weeks it died. I believe it was a sign of a bad card.

Are you running renamed FurMark in 8.8 (e.g., renamed.exe) because it has been nerfed in the driver to not push the cards as much...

Otherwise, the OC looks awesome

[dnottis]

Looks like card #2 died today. When they fail, they just won't post. Playing COD4 last night for awhile, temps were about 38 core, 52 vrm. Shutdown. Today it wouldn't start. Swapped GPUs, posted fine.

Second GPU dead in 2 months. I'm done with ATI. Moving to Nvidia at this point. The first one died at stock speeds. The amount of spontaneous death with these cards is frightening, in fact is SDS - Spontaneous Death Syndrome.

[SNiiPE_DoGG]

Looks like card #2 died today. When they fail, they just won't post. Playing COD4 last night for awhile, temps were about 38 core, 52 vrm. Shutdown. Today it wouldn't start. Swapped GPUs, posted fine.

Second GPU dead in 2 months. I'm done with ATI. Moving to Nvidia at this point. The first one died at stock speeds. The amount of spontaneous death with these cards is frightening, in fact is SDS - Spontaneous Death Syndrome.

bad bios flash? you arent being very clear on the cause of the dead card, cards don't just "die" like that, its got to be something else causing it. PSU, Mobo?

[dnottis]

bad bios flash? you arent being very clear on the cause of the dead card, cards don't just "die" like that, its got to be something else causing it. PSU, Mobo?

Just died. Was fine last night. Shutdown the PC. Today no boot. Swapped out cards, card is dead. Same thing happened with my first card. Google about dead HD4870s, they love to give up the ghost randomly. They do die just like that, I've seen it twice now.

I'm done with ATI for now, 2 is enough.

[Kondik]

Just died. Was fine last night. Shutdown the PC. Today no boot. Swapped out cards, card is dead. Same thing happened with my first card. Google about dead HD4870s, they love to give up the ghost randomly. They do die just like that, I've seen it twice now.

I'm done with ATI for now, 2 is enough.

As stargate125645 would state , it's all nVidia's fault. I bought two sapphire Rads month ago with my new rig so I installed anew and man I had 12 or 13 days of red driver hell and 2 Format's of Vista x64 , never had those problems but that Crossfire hated me so much. And at the last day one of the R4870 died [ Accelero S1 rev 2 + VRM"s cooled extremly well ] , that was the last thing , limit reached, packed them returned took my mony bought GTX280 , no problems since them, maybe I was borned without the ATI.dll that the Silicium god gave us ...

[Blkout]

Looks like card #2 died today. When they fail, they just won't post. Playing COD4 last night for awhile, temps were about 38 core, 52 vrm. Shutdown. Today it wouldn't start. Swapped GPUs, posted fine.

Second GPU dead in 2 months. I'm done with ATI. Moving to Nvidia at this point. The first one died at stock speeds. The amount of spontaneous death with these cards is frightening, in fact is SDS - Spontaneous Death Syndrome.

For what it's worth, you did modify your card. Most other people running the stock cooler haven't had this issue. I don't think it's really fair to point the finger at ATI when you clearly voided your warranty and modifed your card.

[dnottis]

For what it's worth, you did modify your card. Most other people running the stock cooler haven't had this issue. I don't think it's really fair to point the finger at ATI when you clearly voided your warranty and modifed your card.

Ya, but the first one died and it was stock. Regardless 2 dead HD4870s added the pile that others have had die doesn't look too good if you ask me.

[dejanh]

Ya, but the first one died and it was stock. Regardless 2 dead HD4870s added the pile that others have had die doesn't look too good if you ask me.

Ya, I had a reference Diamond HD 4870 die on me as well couple of days ago (5 maybe?). Bought it from NCIX, came home, installed it, installed the drivers, rebooted, loaded up 3DMark06 ran a couple of tests, rebooted to go to Ubuntu to install the drivers and I was greeted with a black screen at POST...card trying to spin up the fan over and over and over, diagnostic LEDs blinking, no video out. Took it back to NCIX 2 hours later, they put it in their system to test it right away...no video...full refund.

I understand that there is a lot of these cards that are being sold, a lot of people that do not have complaints, but I have just been having s*** luck with them...

[0ut0fstep]

Ive now modded 4 of the visiontek cards 2 for myself and 2 for a buddy. using mcw60 block and just cutting the stock cooler down to cool the VRM's. none of these cards have crashed while running furmark ive reached the 120c mark after 130seconds of stability testing on furmark. but the cards are all still solid. im thinking its a manufacturing flaw on someone's part and its not ati's imo because at stock setting with stock cooler the VRM's run plenty within spec using a computer as it should be. Diamond HIS powercolor etc.. if these cards are failing visiontek is doing something they aren't. and yes I have renamed furmark so i get double the frames and stress it as much as possible.

[dnottis]

Ive now modded 4 of the visiontek cards 2 for myself and 2 for a buddy. using mcw60 block and just cutting the stock cooler down to cool the VRM's. none of these cards have crashed while running furmark ive reached the 120c mark after 130seconds of stability testing on furmark. but the cards are all still solid. im thinking its a manufacturing flaw on someone's part and its not ati's imo because at stock setting with stock cooler the VRM's run plenty within spec using a computer as it should be. Diamond HIS powercolor etc.. if these cards are failing visiontek is doing something they aren't. and yes I have renamed furmark so i get double the frames and stress it as much as possible.

Mine have never died while pushing them. They die when trying to come on. Strange. It was fine all day yesterday was playing some COD4, RS vegas 2. Shut her down. Today the card was dead, no more posty, ever. Just like that. My other one did the same thing, just up and randomly died. You better knock on wood so you dont have them all die on you tomorrow

Picked up the GTX 260 from the FS forums here today.

[dejanh]

Ive now modded 4 of the visiontek cards 2 for myself and 2 for a buddy. using mcw60 block and just cutting the stock cooler down to cool the VRM's. none of these cards have crashed while running furmark ive reached the 120c mark after 130seconds of stability testing on furmark. but the cards are all still solid. im thinking its a manufacturing flaw on someone's part and its not ati's imo because at stock setting with stock cooler the VRM's run plenty within spec using a computer as it should be. Diamond HIS powercolor etc.. if these cards are failing visiontek is doing something they aren't. and yes I have renamed furmark so i get double the frames and stress it as much as possible.

From the people I have had an opportunity to talk to so far, I have heard the least complaints about Visiontek cards. Kind of funny...maybe they are more stringent in ensuring no faulty cards come out because of the lifetime warranty

[Boyu]

wow 2 card died, something wrong with your system maybe?

[xMrBunglex]

wow 2 card died, something wrong with your system maybe?

makes you wonder. one dead card is bad luck. two dead cards often means you're doing something wrong.

[dnottis]

Nope, popped in another card no problems. Had my 8800GTX in there before the HD4870 for 2 years without a problem, all the same hardware. Honestly the HD4870 seems like the Wolfdale cores - they just can't play hard or they die. The first card had issues from the day I got it. Would just lose signal and shut down while the PC was still on. It finally died on me without overclocking or doing anything much with it.

The VRM temps are very high which I feel leads to a quick failure. Also the DDR5 runs hot on these cards. I had a fan blowing over them and the ram sinks still got pretty warm, much warmer than my 8800GTX ram sinks ever did. I believe the high temps of the ram and VRMs attribute to the high rate of failure with these cards. The failure rate for the HD4870 is pretty high if you look into it. The card wouldn't let the PC post (beep) so it wasn't even an issue with drivers or overclocks as the card boots up in 2d settings. Once I swapped for an X850XT PE PCI-e it fired right up. Tried a few things, spent the better part of two hours messing with it, nothing, no beeps, no post. Was fine last night and I was playing COD4 for about 2 hours.

What really bothers me is having to change GPU blocks out for the GTX 260. I just rebuilt the water cooling setup about 2 weeks ago. grr.

Anyone looking to buy an HD4870 - really consider the reliability factor with these cards. Might be better to stick with HD4850, lower VRM heat and DDR3 seem to be a better bet.

[xMrBunglex]

i don't know, mang. it's hot as hell where i live and i punish my cards. no reliability issues over here.

[dnottis]

i don't know, mang. it's hot as hell where i live and i punish my cards. no reliability issues over here.

Well good luck with them. I dont wanna see anyone's card fail. My buddy just bought one too and he's really nervous about it now as well. Hopefully all the bad ones were sent to ZZF where both of mine came from.

[Kondik]

Well, I bought the R4870 from sapphire , the next day two acceleros S1 rev 2 , the temps went magically down , I was so happy , but still those driver issues pissed me a bit off but still I benched a bit and allso played Crysis and Devil May Cry 4, still the drivers Installation and uninstalation process after a couple of tries killed my OS [ BSOD fest ] , a format helped until I started to install the drivers again. There simply weren't drivers that worked for me in benchmarks and Games so I stayed with the 8.7 beta and one lovely day one of the R4870 died , noticed it in GPU-Z , just one GPU , tried that R4870 alone, no post ... VGA dead , the temps on core were good VRM's were cooled enough [ never went over 70 ] , RAMSinks allso. The cards were OCed to 775 / 1000 Still the card died. I returned the stock cooler on it, plugged out the second card and returned it to the shop got refund and bought another card. As I said before, maybe there is a magic behind the driver installation of ATI cards which I didn't follow.

But I still think that the R4 series is one big step forward , it just lacks propper driver support, I had issues with every single driver I tried from Omegas to DNA's to official or beta drivers. Now my GTX worked like a charm, installed one drivers OC'd it never looked back, If this happend to me with the R4870's I'd stay with them.

[twwen2]

Has anyone had any whistling caps on their 4870?

I've been running mine for a few weeks now on the ASUS bios (790/925 i think), everything is fine except when i load the crysis demo it starts making funny noises!

Even during the intro videos, it makes this high-pitched squealing noise, stops when the screen changes (in the intro screens) then starts again. It is constantly going whilst navigating the in-game menu until it crashes (yes the menu crashes it).

Now i'm not really fussed cos i don't play crysis (i only have the demo), but i'm a little perplexed none-the-less.

I'm running Cat 8.7 on XP Home SP3.

[|Oc|REvO_HsArC]

Has anyone had any whistling caps on their 4870?

I've been running mine for a few weeks now on the ASUS bios (790/925 i think), everything is fine except when i load the crysis demo it starts making funny noises!

Even during the intro videos, it makes this high-pitched squealing noise, stops when the screen changes (in the intro screens) then starts again. It is constantly going whilst navigating the in-game menu until it crashes (yes the menu crashes it).

Now i'm not really fussed cos i don't play crysis (i only have the demo), but i'm a little perplexed none-the-less.

I'm running Cat 8.7 on XP Home SP3.

Mine whistles in heavy load but doesn't crash...
But yours cant be that loud as I only notice the little noise with my ear next to the case and because my pc is very very silent.

how is yours?

[Shogan]

Has anyone had any whistling caps on their 4870?

I've been running mine for a few weeks now on the ASUS bios (790/925 i think), everything is fine except when i load the crysis demo it starts making funny noises!

Even during the intro videos, it makes this high-pitched squealing noise, stops when the screen changes (in the intro screens) then starts again. It is constantly going whilst navigating the in-game menu until it crashes (yes the menu crashes it).

Now i'm not really fussed cos i don't play crysis (i only have the demo), but i'm a little perplexed none-the-less.

I'm running Cat 8.7 on XP Home SP3.

I get the exact same. Well mine happens in Furmark, and other 3D apps. Some apps do it some don't. Seems to be the way in which the app stresses the card. quite annoying.

[twwen2]

I get the exact same. Well mine happens in Furmark, and other 3D apps. Some apps do it some don't. Seems to be the way in which the app stresses the card. quite annoying.

Interesting. The wierd thing is so far; it only happens in the crysis demo. I've been playing C&C3 for hours on end and it hasn't happened once. It's really quite loud!

[dnottis]

Well, I bought the R4870 from sapphire , the next day two acceleros S1 rev 2 , the temps went magically down , I was so happy , but still those driver issues pissed me a bit off but still I benched a bit and allso played Crysis and Devil May Cry 4, still the drivers Installation and uninstalation process after a couple of tries killed my OS [ BSOD fest ] , a format helped until I started to install the drivers again. There simply weren't drivers that worked for me in benchmarks and Games so I stayed with the 8.7 beta and one lovely day one of the R4870 died , noticed it in GPU-Z , just one GPU , tried that R4870 alone, no post ... VGA dead , the temps on core were good VRM's were cooled enough [ never went over 70 ] , RAMSinks allso. The cards were OCed to 775 / 1000 Still the card died. I returned the stock cooler on it, plugged out the second card and returned it to the shop got refund and bought another card. As I said before, maybe there is a magic behind the driver installation of ATI cards which I didn't follow.

But I still think that the R4 series is one big step forward , it just lacks propper driver support, I had issues with every single driver I tried from Omegas to DNA's to official or beta drivers. Now my GTX worked like a charm, installed one drivers OC'd it never looked back, If this happend to me with the R4870's I'd stay with them.

I'm inclined to think, the underpowered / overheated VRMs OR DDR5 are the culprit for the HD4870 deaths. These deaths aren't isolated, a quick google search turned up a bunch of folks with the "sudden death syndrome" of the HD4870s. Mine have each lasted about 3 weeks and always died while powering on. Never while pushing, benching or gaming which strikes me as strange.

Someone I know that understands more about electronics thinks it may be due to the heating / cooling of the card. While gaming the VRMs would reach 80+c, then shutting off my machine in an air conditioned room it would cool to about 70F. Its like having a computer run 24/7 for months, then shutting it off and having it not come back on again. The constant swing from very hot to very cold in my case may have been the death of them. Using all the same hardware I had my 8800GTX OC'd and water cooled for about 2 years without a single hiccup or failed boot.

The other possibility is the failure of DDR5. There are documented cases of degredation while overclocked. I know on both of my cards the ram got very hot even with airflow and heatsinks on them.

The HD4870 experience has been mixed. Some games ran great, others had some performance drops (COD4 with alot of smoke). After 2 failures I'm tossing in the towel however. I don't believe the HD4850 suffer the same fate because the VRMs don't get anywhere near as hot and their use of DDR3.

The other deaths mentioned here aren't any different that many of the other reported deaths. They just up and die.

[0ut0fstep]

Mine have never died while pushing them. They die when trying to come on. Strange. It was fine all day yesterday was playing some COD4, RS vegas 2. Shut her down. Today the card was dead, no more posty, ever. Just like that. My other one did the same thing, just up and randomly died. You better knock on wood so you dont have them all die on you tomorrow

Picked up the GTX 260 from the FS forums here today.

Sounds to me like a power issue. Electronics that arent getting enough juice voltage/amperage wise tend to die when you shut them off and try to power them on again. Just to be on the safe side I would take your power supply and put it on a voltage tester or test with multimeter make sure you are within spec. Failing electronics are like light bulbs they never die when they are on they die when you hit that power button and give it that last little surge. It coulda been the cards but after 2 dying on you the same exact way I would check your power make sure its clean to be safe man.