I just received my gainward GTX 480, was the high temps a lie?
- Thread starter Ryeong
- Start date
Ryeong
New Member
huh? I made a thead Entitled "I just recieved" and i was speaking of what i see in front of me. I don't see any other people here that have 480 yet, so how can i reflect on something only i can see per say?
If someone else has the card here and see a different result, then i am more than willing to discuss that etc.
CdnAudiophile
Member
So you run your card at 75% all the time? From all the reviews the 5970 is the better choice. It has more potential and takes less watts while producing less heat. Fermi is good at heavy tesalation but is medicore with everything else. In DX10 games the 5870 beats it and the 5970 gets about 30% more FPS. In DX11 game fermi is 5 to 10% faster than a 5870 and still falls short to a 5970.
Ryeong
New Member
Not really. 5970 is a Dual gpu and reviews states that the card has micro-stuttering etc whilst i've not seen a single stuttering on my GTX 480 now. GTX 480 is a single GPU and is so close, and even better than 5970 in some tesselation-instances. I for instance could run Metro 2033 very high, DX11, 1080p with 2xMSAA, DOS, and 16xAF (didn't really see much difference with AF ON and OFF so i might disable it later) without any problem. average fps was 45-50, min 34-36. Two of these in sli is a good choice what so ever you may say. Like i've said earlier, it's up to each and other what their preferences are and do not try to standardise 5970 becasue it's not ALL that good to be a dual gpu. Less watttage sure, but not less heat. MY 480 temps are in most cases was less than what i've seen in reviews, yet i lost much respect from reviewers after seing the truth yesterday.
Correction: GTX 480 beats 5870 in all DX10 games with the exeption of Aliens VS predators. on average 5-10% which isn't that much. Yet, in DX11 games i saw 20% increase in Just cause 2 compared to the 5870 benches. Overclocked i saw even more, so keep that in mind. This is something i see right in front of me btw.
In heaven i saw over 100% increase over 5870 when i used Extreme-tesselation mode and got an average fps of 40, which is better than 5970! On Normal mode i got 60.3 fps the first round and 65 the second. So this SINGLE GPU-card is surperior to 5870. Also, let's talk about Physx. I saw some nice boost there, in crysis i could max the game with VERY high, 1080p, 4xAA and have average 55-58 fps in single player (thus this seemed to vary a lot but never went below 39, unless when the auto save appeared) and over 80 in multiplayer on aveage.
Yet, i've proven the reviewers wrong. The 75% noise was less than my GTX 275 on 60% which is a big plus. Yet, 100% was very noisy and something i just can't use unless i want to test the max OC on air.
Ask me if you want to know anything because i have the card now. I see that it needs some drivers here'n there but it's a living beast!
Max temps i saw after playing Metro 2033 for 6 hours active playing plus 2 hours just being idle in-game (that's 8 hours) was 72c overclocked from 700 core, 1401 Shader, 1848 Memory. To a huge Core 810MHz, Shader 1621MHz, Memory 4000MHz.
Took 30 sec to coold down to 33-35c after exiting.
Last edited:
just a noob
Well-Known Member
Of course you're not going to see micro stutter with a single gpu, it would be impossible, since it lacks another gpu to "talk" to during the process of rendering frames.
Ryeong
New Member
True, but i was thinking about in-game loading when you enter a new zone without any loading screen.
funkysnair
VIP Member
why are you showing realtemp with your cpu temps???
show me the gpu temps under full load and idle...
show me the gpu temps under full load and idle...
The Chad
New Member
Dude, come on. Talk about something else other than tesselation and theat heaven benchmark.
This thread you made about your gpu temps, talk about that instead now. As it was said earlier, if you can run 3Dmark06 and heaven with HWmonitor going. I'm also interested in seeing the min.current.max.
I read some reviews a while ago and lots of them complained about heat. Do you think its dependent on the brand? (Gainward, HIS, Sapphire). And are you ever looking into going Sli or watercooling that bad body?
This thread you made about your gpu temps, talk about that instead now. As it was said earlier, if you can run 3Dmark06 and heaven with HWmonitor going. I'm also interested in seeing the min.current.max.
I read some reviews a while ago and lots of them complained about heat. Do you think its dependent on the brand? (Gainward, HIS, Sapphire). And are you ever looking into going Sli or watercooling that bad body?
Ryeong
New Member
Overclocked from stock 700 core, 1401 Shader, 1848 (2896) Memory. To a huge Core 810MHz, Shader 1621MHz, Memory 2000 (4000)MHz.
33c idle.
72c after 8 hours of metro 2033 gameplay.
80c with 100% stress test on 15-17 hours. (over night)
I'll sli it later, and i don't see any reason why i would need to water cool it. Maybe i'll do it if they add some GPU-coolers the places i buy stuff from so i can overclock even more.
Edit: all cards are default, no brand offer any costum builds yet.. they will probably do so later on. (With the expeption of Evga GTX 480 superclocked and their costum water-block on the card)
bomberboysk
Active Member
Just use furmark and post some screenshots, it has a built in temperature monitor and is about as high of a load as you can put on your gpu.
Ryeong
New Member
I'll do that. The sad thing about furmark is the fact that it pushes GPU's far beyond what they'll ever hit in terms of temperature in gaming.
funkysnair
VIP Member
same as prime95 for cpu's... i doubt i'd get my 4cores (8 threads) running 100% for a few hours but its a standard that people use to test out the cooling....
zombine210
New Member
ummm yeah... do you normally keep this room temp?
that would definitely make a diff with the card temps.
this is too cold for normal people though, my room temp is usually 24/26 when i game.
would be interested in the 3dmark test, will you be posting that any time soon?
thnx
Ryeong
New Member
I didn't bother to wait longer than 25 min, but i doubt it would increase much more because it kept going up to 68 and down to 67. MAX temp 69c after 25 min.. I did some changes in my case and i belive it helped the air flow a lot. Also, the heat from GTX 480 is being lead out from the GPU on the other side (back of my case) so it doesn't really affect my case temp.
Clocks
The test was done with 75% fan speed and OVERCLOCKED from the stock To a huge Core 810MHz, Shader 1621MHz, Memory 2000 (4000)MHz. See Pic Beneath:
Additional Notes:
Current room temp: 25c (that's 5c more than yesterday)
Fan Speed: 75% NOT AUTO (i'll test using 100% fan speed another time)
I did make some changes to my case's air-flow (changed the direction of some fans and this alone decreased the 480's temps by almost 10c)
After 25 Min (the temp kept going up to 68 and down to 67 all time time, so i doubt it would go much higher if i let it stay for any longer:
After 15 min:
funkysnair
VIP Member
give us the temps on auto, dam i could have cranked up my 4850 fan up to 75% when i had it and said it run cool when i know on auto it ran dam hot!
Ryeong
New Member
The problem by doing so lies in the usage of one spesific tool. I use EXPERTool from Gainward, whilst other tools such as EVGA's and so on might not cranck up the fan speed the same way my tool does. I tried Auto and it starts on 30% and when i hit 70c it just doesn't increase to more than 40% at MAX, which makes me wonder if there is something wrong with my tool or if that's normal because it focus on the thermal themp (105c.)
I had the same problems with my Gainward GTX 275's in sli. On Auto they ran way too hot without increasing the fan speed much. Not good thing.. I only use an Updated version of EXPERTools (CD came with the card) for my Gainward GTX 480.
The reason why i keep it at 75% is because all the other fans in my case (even the fans to my H50-cooler) makes more noise compared to my 480. Also, Gainward GTX 275 made more noise on 60% than 75% on my 480 which i'm pleased with.
I have, in the previous days, made some benchmarks and are prepearing to make an objective overview of my gainward gtx 480 that ONLY focus on pure results rather than my personal opinions. I will include everything then.. benches, temps, pics.. everything.
Last edited:
CdnAudiophile
Member
So you run your card at 75% all the time and then tell us it doesn't run hot? Noone in there right mind would have there cards fan at almost full blast all the time. Way too loud for the normal user. Can you do some test's on stock speeds with the fan on auto. Just like someone that bought the card, installed the drivers and started playing their game. (average user)
Ryeong
New Member
75% fan speed was not noticable due to the other components in my case being Louder (Antec 900) than the GPU. Also, the Noise was less noticeable on 75% fan speed compared to GTX 275 on 60%.
Edit: i'm currently making an objective review. If something isn't included in the review, i'll add any request later on.