Simple, get the Club3D card. The PowerColor 4870 PCS+ has problems (see here for details http://en.expreview.com/2008/08/23/p...-issues-check/)Quote:
Originally Posted by Break16
Printable View
Simple, get the Club3D card. The PowerColor 4870 PCS+ has problems (see here for details http://en.expreview.com/2008/08/23/p...-issues-check/)Quote:
Originally Posted by Break16
@screenshots and everything regarding this new bios. Could I add one more request to those posting screenshots? Could you PLEASE post screenshot of GPU-Z "sensors" tab as well, under load ofcourse.. So perhaps when having atitool artifact free for 1 hour, open one GPUZ with first tab, and one with sensors, so we can see all GPU/VRM temperatures and current amps.. Thanks! :)
Btw, I did try this BIOS, and it really ups the voltages. But since I'm CPU limited I've just given it a try, measured real volts to confirm it wa indeed raised, and that's it. Multimeter showed around +0,02V under load vs stock reference BIOS (so not stock XOC, but reference one..though they probably have same volts :) )
Never mind, it's still from official source.. Though it would be nicer if it was from ATi themselves :D ;)Quote:
Originally Posted by IRONBMANwell i recived many emails from Diamond MM and also Ati tec center about the problems i was having with the card, I finaly recived a link to the new bios fron tech02@diamondmm.com so im not sure if it was actualy from ati sory:( was my mistake saying "from ATI".Quote:
Originally Posted by LuxZg
Ah, if these cards and their BIOS were from ATI themselves 4870's would all clock into the 1000s. The days of quality building like that are long gone... :para:Quote:
Originally Posted by LuxZg
@ dejanh the item that steve is talking about come out of guelph ont, his name is andrew and he is a genious:D hear is a link to the card in question
http://www.extreme-pc.ca/showproduct...oductid=372404
i hope i did not brake eny rules posting this link.
I have three of his cards and thay all are rock soild:up:
This BIOS in no way helps my memory overclock, 1100 its unstable, currently testing at 900/850 and it seems its helped the core quite abit, the memory Is stable around 975 I think.
Here we go gents (and maybe some ladies)...time to stir up some s*** :)
I've been posting for weeks now how it looks like there are major issues with 4870 VRM cooling and/or the overall power design...some listened, some did not...
Finally, here is some official confirmation of VRM issues...sadly ATI just "glazes" over this by hacking the driver and not fixing a problem...
http://en.expreview.com/2008/08/26/a...it-run-slower/
Now again, I will care to reiterate over all of the threads that I have been posting in related to this...
This thread is a pretty good collection of all the places I have posted about this already (see first post)... http://forums.tweaktown.com/f31/diam...edition-27348/
Let's hear some opinions!
So basically the 4870 is ATi's version of the 7900GT from NV?
I got the modded bios on, and with Accelero S1 in furmark I dont hit 100 degrees now so I think the VRM's are safe now.
They are talking about poor cooling, which is true. I've upped my fan speed to 75% (which drowns out my annoying roommate). I don't see any proof of the actual VRM design being faulty. Cooling issues are easily fixed, crap electrical design engineering is a lot harder to fix.Quote:
Originally Posted by dejanh
I wouldn't call it a major issue or a design flaw. At stock, the HD4870 can run any game without crashing or dying (at least the ATI reference design, OC version are another story).Quote:
Originally Posted by dejanh
Furmark is just too demanding. It's like runing your car into first gear, pedal to the metal for 10 km and when the engine blows up saying "See, I told you, this car cooling is faulty" :shrug:
That being said, I don't really like the fact that ATI quietly "fixes" furmark issue with a kind of cheat, but if it stops people from burning their card, it's not that bad. Anyways, I don't play furmark so I couldn't care less ... :rolleyes:
I've heared that original Diamond download for that XOC bios is gone so here is reupload on rapidshare.
If it gets deleted, just PM me and I'll upload again..
http://rapidshare.com/files/14071884...00MHz.zip.html
ZIP contains untouched exe from their servers, and I've just added txt file in ZIP, which contains this thread URL, and copies of original post, my post asking if it was ATI or Diamond source, and IRONBMAN's reply confirming it's from Diamond Multimedia..
Is the same BIOS as Ironman uploaded? I'm running that one.
Yes, this is same one, just reuploaded because obviously DiamondMM didn't like everyone downloading it, so they've pulled the file :) If you click on link in Ironbman's post you'll just get 404 error ;)
Wow, I just flashed mine w/ that BIOS...825/1100 just played 60 minutes in Crysis without a hiccup. With my stock BIOS, 1025 was the max mem setting. I'm not saying this 1100MHz setting is stable as the only test I've done is an hour of Crysis, but I'm impressed. I couldn't do 30 minutes before the flash at 1100MHz. 825MHz was my previous core clock max w/ the stock BIOS also -- once I find max stable mem I will test max core clock.
Put the new BIOS XOC, 795/1100Mhz tested, but the ATITOOL accuses artifacts.
If I put more tension by RBE advance something?
Let speeds card in 795/4400Mhz (Stable) has a problem?
It can reduce the useful life of GDDR5?
Cold cards.
???????????????
ATT probably does not work well with the new cards. Try Furmark, but rename the .exe to something else if you are using Cat 8.8.Quote:
Originally Posted by 8800's@over
Nobody knows if it will reduce the life of the GDDR5 or not. These cards have not been out for long enough.Quote:
Originally Posted by 8800's@over
Wow, you actually understood his first question? I tried, but got nothing :pQuote:
Originally Posted by Super Nade
In terms of reducing the life of GDDR5, running it at higher frequency but not changing the supply voltage should be safe. If however you are overvolting the memory and running it at higher frequencies then you are reducing the life span.
Well, I finished working w/ my card today -- 3 hours of Crysis, no problems. Can't get any additional core speed but was able to take the mem from 1025 to 1100 with this BIOS. Or atleast no lockups so far at 825/1100. I haven't tried pushing it beyond 1100 and probably won't. I feel pretty lucky having gone from 1025 to 1100 just from a flash.
oo, i was thinking of buying that one, glad i didnt now.Quote:
Originally Posted by dejanh
I can run my sapphire 4870 with this bios up to 895/1150 in 3dmark01 atleast. Over 895 on core is not stable at all. Without vmods ofc
I tested this BIOS against my original VisionTek BIOS with my DMM and it gives the same load (~1.265V), so I'm not getting the voltage boost from it. I also tried some different higher voltages in RBE and it didn't seem to affect it either. There doesn't seem to be a way to increase voltages beyond hard mods. Undervolting is great though, I run 2D at 1.083V very well :).
Mr K6 - you've used NEW XOC BIOS? How the hell did I than measure +0,02V under load on my Club3D card (reference design)
Also, have you tested under load? Because I get over 1,3V under load with overclocked card (like 800/1100 in this BIOS) even with stock BIOS, and with XOC It's smthing like 1.33V. Load was made by benching Crysis on 1680x1050 High/very high, and measuring on VDDC measure spot shown in 4870 voltmod thread..
Good question, but I have no idea. I'm using the same area (highlighted in blue/purple in the thread). Generally I get in the 1.265V range with any BIOS (also tested ASUS TOP BIOS for the hell of it). Overclocking changes the voltage by about 0.001-0.003V at most, nothing much. Also, I never got increased overclocking capability from the XOC BIOS (I used the one you reposted a bit earlier).Quote:
Originally Posted by LuxZg
Weird, really weird. I'm home now, and can copy my measurements exactly:
in 2d 500/1100 - 1,302V; in 3D 795/1100 - 1,331V-1,334V (jumping around as bench moves around)
For default BIOS, I had these:
in 2d 500/900 - 1,283V; in 3D 750/900 - 1,305V-1,308V
OCed in 3D 790/1100 - 1,307V-1,312V (so matches yours 0,001-0,003V difference from overclocking)
Now, it means I had up to 1,312V with stock bios when overclocked, and on almost same clocks (just 5MHz GPU, that shouldn't change a thing) I get up to 1,334V, so I got +0,022V from this BIOS. Also, my Catalyst showed unlocked Overdrive with this bios, and increased clocks as well.
Have you changed or modified this BIOS in any way before flashing? And are you sure it was actually FLASHED to your card? Perhaps it exited flash operation if you did not use FORCE option for flashing, because of different card IDs? You know you have to run ATiWinFlash in command prompt to FORCE flash any BIOS?
You would be certain if you would flash it by their own flash app by just running exe, not extracting and flashing manualy..
EDIT: If you DID flash correctly, than you can "download" your new bios by GPU-Z, and open it in RBE. If you don't see any changes in BIOS time/date/version and/or clocks, than you haven't flashed at all.. and that would be only explanation at this moment (to me at least :) )
Same here.
I did the flash from command line, but I've gained nothing with this BIOS (I had to edit it before to 795/1000, 1100Mhz is too much for my memory, it seems). Asus TOP and Diamond XOC BIOS both run Crysis @ 860/1010 (Autotune), but ATiTool gets corrupted with anything above 810Mhz Core.
I don't have a DMM at hand, but RivaTuner and GPUz both read "1.26V" in 2D and 3D. I don't know if I should trust those readings, can you post a pic of Rivatuner Hardware Monitoring while runing 3D?
I will flash the XOC BIOS as is, maybe something goes wrong when I underclock the memory.
Sorry for my English...