does furmark up the vrm temps rediculously high as with the 4870/90?
Printable View
does furmark up the vrm temps rediculously high as with the 4870/90?
Damm I want to see OC when people get their waterblocks for these :)
Btw do you think its a bad idear to get the MCW60 and passive mem cooling compared to fullblock when my CM ATCS 840 only have one fan blowing air in? Like do the mem chips get really hot on this board?
i have gpu only + little coolers atm...
memchips do get hot to the touch without active cooling but not hotter as they were on the 48xx
about furmark: yes it burns the hell out of the cards again. with "xtreme burning" mode i get ~80°C @ 925/1300 with stock voltage on the vrm's after a minute, whereas 2h of gaming will stay at 60°C max. power consupmtion gets up to 74A, whereas gaming or atitool would use 50A maximum xD
Did not test with stock cooler, would be nice to have a comparison....
I love Afterburner(with g15 support and small interface will be the ultimate tool)
http://img183.imageshack.us/img183/2...870uruk.th.jpg
Last GPU-Z doesn´t read the overclock speeds :shrug:
i think swiftech will release a nice passive ram and vrm companion for the mcw60, so i would recommend that combination, because gpu-only coolers normally provide better temperatures and last longer then one card generation. i use my heatkiller gpu-x since the geforce 7800^^ saves a lot of money and the temperatures are awesome...
also i found that for most cards gpu-only+some passive coolers and some thermal adhesive is enough really... if u like or dont care about the looks that is..
Anyone know the max voltage for core or Ovp/OCP limit? Or stock voltages for the 5850 / 5870? I hear it is 5850: 0.84v idle, 1.09v load and 5870: 0.95v idle, 1.15v load. So far, two people in this thread have reported 5870 load voltage to be 1.125v and 1.162v...
Also check out these overdrive limits:shocked::
http://i38.tinypic.com/35dd939.jpg
1500/1500 :eek:
My CCC stops in 1200 for core and 1400 for ram.
any of u guys could post what their vrm temps are like under load?
trying to find out whats safe and whats not as well :)
http://www.xtremesystems.org/forums/...6&postcount=43
here you go, this bios is the 1200/1400 one
just crashed crysis with 1000/1300 @ 1,25v, lets try 1,275 :)
what is very nice: msi afterburner writes the voltages into the driver config if u select "apply oc at startup", such that only the 3d profile vgpu is upped even without afterburner being loaded... neat!
Hi oese. I have now flashed my xfx to asus and it worked perfect. Now the 2dclock works again with oc and voltage. But I want to know how high voltage is safe for 24/7 ? Is 1.225v with 1000 core and 55c max temp ok for 24/7? Or can I use up to 1.275v for 24/7 if the temps are ok and no artifacts? Oese, do you use that high volt for 24/7 or just for benchmarks ?
/wekktor
any asus 5850 bios yet?
I can link you to the older topics that discuss OCP & OVP for rv770. I'm still waiting for someone to tell us the point where OCP starts for cypress. (and also to measure default idle / load with dmm :p:)
You can find 4800 series discusson of OCP here: http://www.xtremesystems.org/forums/...d.php?t=190990Quote:
Disclaimer:
Performing these voltmods will void any and all warranties your video card may have.
You - and only you - are responsible for any damage caused indirectly (or directly) by these modifications.
For example OCP on HD4870x2 is 1.37v. Without hard modding & soldering on your card, the card won't let you software modify each gpu's voltage higher than 1.37v. That interference is called triggering OCP. It's not something found only on 4870's, its in all NV 8000, 9000, Gtx series, ati 4000, 5000 etc. It is the card's method of protecting itself from voltage surges, spikes, and dumbos illegally tampering with vgpu :D. The extreme oc'ers even make PCB modifications that bypass OCP & OVP (extremely dangerous) in order to push even more voltage into the core.
Like this: (http://www.xtremesystems.org/forums/...7&postcount=81 and http://www.xtremesystems.org/forums/...7&postcount=87).
Those extreme guys (kingpin, hipro, & shamino and the likes) are aiming for 1.3ghz clocks on LN2 cooled gpus. Most modern gpus have over-clock protection, over-current protection, and over-voltage protection limits. These limits trigger when internal loads or temperatures are detected over the limit. Also it has been discussed that 5800's have hardware level throttling for the gpu's core speed and the boards voltage regulators (something not present in the 4000 generation - remember the furmark VRM overheating thing) that, in addition to ovp & ocp, will prevent the card from hurting itself (not to mention the memory has ecc & is protected as well). Knowing about all these self-protections mechanisms may make you a little more care-free about overclocking, but you can still KILL YOUR GPU. I've yet to hardmod a vga and know only a little about OCP, and don't want to encourage this type of modding (even though it's fun) because I am not responsible for any risks you make based on reading any of this. Maybe someone like K404 or Largon can explain OCP in more accurate detail, or you can try searching through the grahpics Vmod section here to learn more about it. Hey, obv someone made a 1500/1500 bios, so 1200/1400 must not have been enough. :rolleyes:
:ROTF:
Stock volts, comfortable 40% fan speed. HD 5850 puts up a very interesting fight :)
http://lab501.ro/forum/attachment.ph...0&d=1254574647
http://lab501.ro/forum/attachment.ph...1&d=1254574647
Just a word of warning people.
do not test your graphics cards with firmmark or atitool
as they wont show errors, the memory controller on these cards has intergrated error checking so even if there was errors it wouldnt be shown on screen.
ATI tool will never show memory errors.
test your overclocks with the frost levels on crysis warhead
aswell as stalker clear sky at the start all maxed out 4xaa16af ectect.
There is issues with furmark and the 5870 cards.
the issues your seeing are core instability artifacts.
dont test with firmark as its a sure fire way of burning out vrms pritty quickly
test useing the most stressfull games.
Side Note* that guys firmark above is at 85c lol i would love to know what your VRM temps are under the hood i bet there 100+
how can you see the vrm temps ?
i plan to use for 24/7
but pls ppl could someone look up his vrm temperatures in the gpu-z sensors tab so we can make some kind of a comparison?
i have 80°C after a minute of furmark @ 925/1300 stock voltage and 75°C when gaming crysis @ 1000/1300 with 1,275v vgpu.
all on small passive heatsinks glued to the vrm with thermal adhesive...
pls someone post your temps either stock cooler or anything else....
for vrm temps use the gpu-z tab "sensors"
VDDC phase 1 to 4 are the vrm temperatures...
thx oese
I will try furmark after I ate and see what vrm temps I get
nice :)