Sapphire HD7950 3GB Vapor-X Review

Sapphire HD 7950 3GB Vapor-X Review

http://img.photobucket.com/albums/v1...psaa271e17.jpg


Introduction

I?ll save you the boring company history lesson, anybody who has been into custom building PCs for more than five minutes will know who Sapphire are and the premium graphics cards they offer. Today we will be looking at the HD7950 Vapor-X, can it live up to all the anticipation? Read on to find out.


Gallery

More in-depth than usual today. There?s been so much hype about the Vapor-X that I felt a little more attention to detail is warranted.

http://img.photobucket.com/albums/v1...ps78157091.jpg

Here we have the card from your standard top-down view, you?ll have to forgive the lack of creativity in the photos today ? I had a lot of things to do. At first glance there?s nothing out of the ordinary to take note of, just your common garden variety custom cooler slapped onto a HD7950. Being a Vapor-X this cooler does use the much touted vapor chamber technology, we shall see what kind of difference to temperatures that makes later.

http://img.photobucket.com/albums/v1...ps77b4aea0.jpg

We can get more insight into the design of the cooler here, 4 nickel plated heatpipes and we can see Sapphire have chosen to cool the memory by using a built-in heatspreader using thermal pads to finish off the contact. I?ve never been a fan of this approach to cooling memory on graphics cards, it just seems inherently lazy and penny-pinching. Neither of which Sapphire should be doing here considering this is supposed to be a premium custom designed graphics card.

http://img.photobucket.com/albums/v1...ps8a2c7dcb.jpg

Perhaps one of the more interesting pictures, we see here this HD7950 is using two 8+8 pin connections instead of the more common 8+6 or 6+6 configurations. It would definitely seem Sapphire have built this card to be overclocked.

http://img.photobucket.com/albums/v1...psd71f5d86.jpg

The back of the card doesn?t tell us too much, but that VRM heatsink looks a bit suspect.

http://img.photobucket.com/albums/v1...ps32bf3967.jpg

Now the front of the card, nothing to see out of the ordinary here people, move along, move along.

With the normal pictures done I decided to delve further into the workings of this card so I removed the cooling. This is what awaited me;

http://img.photobucket.com/albums/v1...ps42c94535.jpg

First, the rear of the card. Soldering is top-notch and it?s nice to see Sapphire have at least used washers in most of the areas that need them... that?s right...... most..... You?ll notice a few little screws that are used on the anti-warp bar, there are no washers there. There is no spoon...

http://img.photobucket.com/albums/v1...ps0bd0eec2.jpg

Finally, the front of the card, VRM heatsink removed an all. Sapphire has certainly heatsinked everything possible and we see the very common Chill controller being used. We should have full access to memory and GPU voltage control. I?ll also take this chance to point out that with the anti-warp bar design Sapphire have chosen to use that it covers a small cluster of VRMs, but Sapphire have taken no measures within the anti-warp bar design (aside from a teeny tiny thermal pad on the underside) to ensure these VRMs get proper cooling.. rather worrying. The VRM heatsink on the back of the card also raised alarm bells for me, the VRM heatsink juts out over the forest of resistors you can see. What you can?t see very well is that these resistors sit considerably higher than the VRMs, as such in order to ?fix? this problem somebody at Sapphire decided to use a dirty fat thermal pad to ?bridge the gap? of the height difference between the VRMs and resistors. I don?t think I need to explain to anyone familiar with thermodynamics what a bad idea that is, hopefully the effect of doing this won?t be as negative as thermodynamics tells us it will be.


System Setup

You should all be familiar with the setup now, but for those that aren?t;

I5 2500k @ 4.5GHz
2x4GB G.Skill Ripjaws @ 2133MHz 10-10-10-15
Asrock Z77 Extreme6 w\ UEFI 2.3
Sapphire HD7950 3GB Vapor-X w\ Catalyst 12.11 Beta11
640GB WD Caviar Black, SATA2 32MB Cache
2x SATA Sony Optiarc DVD-RW
XonarDX 7.1
850w Corsair HX PSU


Benchmarks

The benchmark session today will be brief due to time restrictions but none the less you will be able to get a very good idea to this cards current and future performance. With that in mind I have chosen to use 3DMark 11 and Unigine Heaven 3.0. For all tests I?ll be using vBIOS ?2? which is the boost vBIOS for 950MHz. Results initially were a bit skewed but I discovered that was due to the GPU frequency bouncing around. Raising the board power limit sorted that out though.

3Dmark11 (performance) stock score:

http://img.photobucket.com/albums/v1...ps5f18d54f.jpg

To be expected performance here for 950/1250 clocks, nothing of interest to note.

Unigine Heaven 3.0 (default settings) stock score:

http://img.photobucket.com/albums/v1...ps55feeb5e.jpg

Now things get a bit complicated. Upon moving to the OC tests I came up against a brick wall. Afterburner couldn?t adjust GPU voltage or memory voltage, so I moved to TRIXX 4.3, which also couldn?t adjust GPU voltage and didn?t have any way to increase memory voltage. So next up I hit Google, I discovered a beta version of TRIXX (4.4), which... you guessed it, also didn?t allow GPU or memory voltage adjustments. None the less I hung in there and came across a modded version of TRIXX 4.4, which does allow voltage change for the memory, but you still can?t adjust GPU voltage. This modded version of TRIXX is the best I could find so it?s what I?ll have to use, the rest of this story I will get into a little later but suffice to say it turned this review into a very long winded affair.

3DM11 (performance) OC (1155 / 1600) score:

http://img.photobucket.com/albums/v1...psf2882325.jpg

Unigine Heaven 3.0 (default settings) OC (1155 / 1600) score:

http://img.photobucket.com/albums/v1...psd06397f1.jpg

Unigine Heaven 3.0 (custom settings) OC (1155 / 1600) score:

Custom settings used:
Tessellation: Extreme
Shaders: High
16x AF
8x AA
Catalyst AI set to ?very high?

http://img.photobucket.com/albums/v1...ps1ce38c77.jpg

Now to fully divulge what I briefly touched upon earlier. The next step of this review was to be to report on VRM temperatures, specifically to address the concerns of the thermodynamic issues I mentioned earlier. Now, I did start this process for all of about four minutes, which is all it took for the VRMs to heat up to a insane 120c as reported by my DT8380 temp gun which is accurate to +/- 2%. At this point I promptly stopped the tests and went about doing some very in-depth homework. If there?s one thing this part of the test did verify is that I was absolutely right to be concerned with how Sapphire has chosen to cool the VRMs. My research lead me to numerous places, I?ll provide links to a few of them to give you a idea of what I come across.

Sapphire Forums
Thinkdigit Forums
anandtech Forums

Those links are just the tip of the iceberg. After doing much reading I decided to get in contact with Sapphire to see what they had to say so I posed some questions for them to answer such as why is it vBIOS 15.25.01 for the Vapor-X has a MVDDC of 1.6v but vBIOS 15.25.99 has a MVDDC of only 1.5v, along with other questions such as the high VRM temps and why is it neither TRIXX 4.3, 4.4, Afterburner or GPU Tweak aren't able to adjust GPU voltage. To be clear about that final bit, you can still adjust the GPU voltage slider and hit the apply button, but the voltage is not applied. After submitting my barrage of questions I sat back and waited... and waited... and waited... and waited. It?s now been some 3 weeks since I posed these questions to Sapphire and I have still not got a reply from them to any one of the questions I asked. This is absolutely unacceptable and furthermore I am flabbergasted that no other reviews for the Vapor-X have addressed these concerning issues, particularly the VRMs heating up to 120c!

As such all I can do now is bring these issues to as many people?s attention as possible in order to hurt Sapphire in the only place they will feel it and perhaps incline them to become helpful ? their sales of the Vapor-X cards. I have even gone the extra mile and used my knowledge of thermodynamics to help people with the overheating VRMs who are stuck with this card.

Here is what you need;

- Dremel with drill bits & chucks
- Metalwork vice
- Jr. hacksaw or electric metalwork saw
- Medium & very fine grit sandpaper
- Selection of small screws, nuts & washers OR some spring loaded pushpins
- TIM or very thin thermal pad
- Small heatsink & thermal tape

Step 1:
Remove the anti-warp bar and apply a small, lightweight, reasonable height heatsink to the small hidden cluster of VRMs.

Step 2:
Now you will want to drill out the brass standoffs from the VRM heatsink, so take a drill bit about the same circumference as the bass standoffs and drill them out, leaving a couple nicely sized holes. At this point also notice the horribly thick thermal pad Sapphire use... this is THE worst thing to do to "fix" a poorly designed heatsink.

Step 3:
Flip the heatsink over, see that bit of plastic attached on the bottom? You need to saw around this. If you are using a Jr. Hacksaw this will take a while, a good 45 mins, be patient and take your time, you really can't afford to balls this part up. If you are lucky enough to have access to a electric metalwork saw you'll have this part done in no time flat, but all the same take your time and be careful.

Step 4:
With that lump of metal cut out all you have to do here is trim down each end of the heatsink where you drilled the holes earlier, leave just 1-2mm of clearance either end. With this done you have now completed making your clearance so the heatsink doesn't interfere with any of the resistors in the VRM area on the graphics card.

Step 5:
Now it?s time to get sanding, use the medium grit to clear off the black paint from the base, then use the fine grit to really even out the base of the heatsink and make it as smooth as possible.

Step 6:
Another easy bit, apply a thin thermal pad or a thin layer of TIM along the entirety of the heatsink base, attach the heatsink to the graphics card with a couple small screws and nuts. Don't be shy about tightening the nuts as much as you can with your fingers, but also use a little common sense here and watch for PCB warping. If you don?t trust yourself with a few screws and nuts, use a couple spring loaded pushpins but be sure the spring can give a reasonable level of pressure so the heatsink doesn?t rock.

So, after that detour we reach;


Conclusions

Naturally, everything I went through is going to be factored in here. Firstly we will look at the PCB build quality itself. The PCB is undeniably built to a high standard and performance is what you would expect it to be from a HD7950, but without a lot of work this is where the positives end.

The Vapor-X cooler the instant you start to OC is ungodly loud at anything above 55%, with some foam pads strategically placed on the cooler if you were to totally disassemble it would quiet it down a lot, but it still won?t negate this next part.. the Vapor-X cooler is just outright not up to any kind of OCing. At the maximum OC I could attain with this card (1155 / 1600, GPU voltage 1.025v memory voltage 1.6v) the core heated up to over 81c in minutes. By contrast fitting a Gelid Icy Vision A cooler temps with the mentioned settings did not exceed 58c and the Icy Vision was significantly quieter even at 100%.

Next is the problem with the VRMs heating up to ridiculous levels, 120c AT STOCK in Heaven 3.0 in the space of less than five minutes is just atrocious. Fortunately if you follow my advice above you can fix this problem but come on this is the very basic of basic thermodynamic principles the fact Sapphire can?t even get this right just makes me wonder what idiot they tasked with designing the cooling for this card and the kind of people they employ in general. The latter of that is particularly questionable when Sapphire stone wall their own customers. Here?s a little hint Sapphire ? you won?t last long in business adopting these practices.

As if to add insult to injury Sapphire seem to actively be ignoring these issues as can be witnessed in their own forums.. no official word from any of their techs, nothing. The smaller issues such as GPU and memory voltage while making a LOT of people unhappy could perhaps be overlooked if it was JUST these issues, but in light of all the other aforementioned issues as well they simply compound the misery and what is a sorry excuse of a product from Sapphire. The card is fundamentally flawed and a product recall should be taking place to fit adequate VRM cooling.

In closing I would generally say avoid this product like you would the bubonic plague, but if you are handy with a soldering iron and have thermodynamic knowledge and want a project, you?ll get one with this 7950 Vapor-X. To anybody else though, avoid.

Final Verdict: Abysmal

I will be sending Sapphire a link to this review, perhaps it will prompt them into being more cooperative toward everybody and we will finally get some things resolved.

People might find THIS interesting. By poking and prodding and making noises in the right places you can get a manufacturers attention, as this is proving :up:

Review looks good, one type, You said "Chill" controller, you added an extra L. ;)

Yeah I know, but I find it hard to deliberately spell things wrong :p: On the plus side I've got Sapphires attention and there is at least movement. We'll see how far it goes.

Ket - nice write up, I have a couple of questions/comments

1 - Is this card a boost card?
2 - I'd be interested to see how the card performs in some other benchmarks, preferably non-synthetic :P
3 - I'd suggest looking into some different bios versions (looks like you're already doing that). Some folks with the Gigabyte cards had a large variance in the success of their overclock depending on which bios version they used (link below)
http://www.overclock.net/t/1313813/n...igabyte-hd7950
4 - Just a comment - I would really like to see VRM temps in actual games. 1600mem clock is very impressive with stock cooling (I think you ran those benchs before the mods, right?). Most 7950s are maxing out at around 1450 on the memory.
5 - What VDDC was your card at when you got to max gpu clocks? (different bios versions provide different vddc as shown in the gigabyte link above)

I just installed my gigabyte 7950 last night so I'm excited to see what I can do with it this weekend :)

1. Yep this card is supposed to be the boost edition. Interestingly in the sapphire forums link I posted further up Sapphire tried to deny this card is supposed to have a OC vBIOS, THEN they tried saying vBIOS 15.25.01 isn't meant for my Vapor-X even though it came pre-flashed on the card and the BIOS PN numbers (except for the internal revision bit, which doesn't matter) are a match for this Vapor-X. (screens below) AND (yes theres a and) tried saying I flashed the wrong vBIOS on it, even though I stated numerous times it came pre-flashed to the board :hm:

http://img.photobucket.com/albums/v1...ps9d34b155.jpg

http://img.photobucket.com/albums/v1...psf7c2f52a.jpg

2. Those results will come once Sapphire get their thumb out their arses and send me a vBIOS with 1.6v memory voltage and a copy of Trixx (incoming v4.6, apparently) that actually allows for GPU voltage adjustment.

3. No need if somebody at Sapphire does their job right, but I'll bare it in mind for if the all else fails (namely, Sapphire fail to come good) scenario arises.

4. In actual games VRM temps hit in the region of 140c with stock clocks. The reason theres no screenshots for it is because every time I waited for the VRMs to peak the system crashed.... not really surprising when the VRMs are hitting 140c or more. (with my mods temps haven't passed 100c in even the most demanding games with the max OC and max Heaven 3.0 VRM temps are about 85c. VRM temp tends to hover around 86c - 90c with max OC gaming) The reason I can hit in excess of 1600MHz on the memory is because the OC vBIOS that came with this card has higher memory voltage, 1.6v vs. 1.5v.

5. Max voltage under real heavy load with max OC sat around 1.018v, which means the card was dynamically adjusting voltage a bit every time I raised board power limits as bone stock voltage was about 0.95v.

I won't let this rest, especially as Sapphire aren't even having the good grace to admit theres been a screw up and until I really shined a light on these issues they were quietly trying to sweep it under the rug instead of just fixing it and ignoring users who noticed the same things I did (I can provide links/images if you want). The problems with the Vapor-X are easily remedied, even the poor VRM cooling if you are a enthusiast. Sapphire certainly shouldn't keep selling this Vapor-X though with the incredibly inadequate VRM cooling it currently has. They should A: redesign the VRM heatsinks for future shipments, and B: offer a recall to customers who have already bought a Vapor-X in order to fit the (much) improved VRM cooling, or at the very least anybody contacting them with a Vapor-X complaining of bad temps bloody well send them the upgraded VRM cooling for them to fit.

Hopefully all of that made sense, I'm in a bit of a hurry atm. You should read through the Sapphire Forum Thread I made, the read... is interesting. Even though I've categorically proved what I assume to be one of the Sapphire techs wrong and caught him BSing numerous times now.. It'll be interesting to see whats said now I've shown the Elpida and Hynix memory don't use different voltages, like they have tried to dupe people into believing. Whoops, I should say "saying". :p: Which has been their excuse for why different Vapor-X vBIOS versions (even ones obviously meant for use with boards using the same memory) having different memory voltages. The real truth is that they have quietly been trying to stop the VRMs from overheating on the Vapor-X cards due to the crap VRM cooling.

I'll keep this short as I'm pushed for time but the main problem of the Vapor-X is that if I use vBIOS1 the Vapor-X is not stable once vRAM usage gets to 1200MB - 1300MB (the card itself crashes, not the driver), but if I change the memory voltage to 1.6v (instead of 1.5v) the Vapor-X is perfectly stable with full vRAM usage, so naturally I'd like the latest vBIOS for the Vapor-X but with memory voltage set to 1.6v to stop the crashing. This issue, as I've also said, is because the memory voltage is too low.. which both Hynix and Elpida's own technical sheets verify.

As yet Sapphire have still not provided me with a vBIOS simply changing memory voltage to 1.6v. I'll keep you all posted of progress, but so far the longer it takes Sapphire to fix these problems, the more damage it will do to their reputation.

If they sell the cards in .eu it would be a class action if people caught on

Most stupid thing is I've done Sapphires job for them, all they have to do is change memory voltage to 1.6v in a vBIOS update, and release Trixx 4.6 which is supposed to allow voltage control of the GPU for Vapor-X and other cards. I shall continue to hit Sapphire hard by spreading word of this abysmal situation to other large tech sites like TPU if they don't take their thumb out of their arses soon.

Keeping people informed, here is the last response I got from the Sapphire tech I'm corresponding with and my reply. Hopefully we will all start to get these issues resolved now.

Quote:

---

Hi Dave :

Thanks a lot for your update.

I will need your help about some information that we will like to duplicate this issue at our side.

Actually , we still not got related information about this kind of testing would crash the board.



1. What applications or 3D games that you used to test the card ?

2. How to check the VRM usage during 1200~1300MB , use GPU-Z ? (We only saw memory usage , is this what you meant ?) any photos can provide to us ?

3. Please show related photos about how to adjusting the memory voltage with TriXX V4.4.0b

4. Please provide us your detail platform information included OS and PSU.

5. Can you try to set to V1.55V at your side and try again if will crash or not ?



Thanks a lot for your help.

---

My reply:

Quote:

---

1. To maximise vRAM usage I have tested with Skyrim using the HD2K texture mod (http://skyrim.nexusmods.com/mods/607) as well as my own HD textures for Stalker: Shadow of Chernobyl (http://www.moddb.com/mods/stalker-twilight-ket-2zero) both games crash when vRAM usage goes above 1200MB - 1300MB. Sometimes it varies a little and won't immediately crash, but it never takes long once vRAM usage gets to the 1200-1300MB region. For clarification, I have tried using Catalyst 12.10, 12.11 Beta11, 13.1 and 13.2 beta drivers, each time doing a thorough cleanup after the normal Catalyst uninstall process with atiman uninstaller 7.0.2 and Driver Fusion 1.1. This problem is most definitely because memory voltage is too low as like I have said, setting the memory voltage to 1.6v completely fixes this problem and I have also used a MSI HD7950 TwinFrozr and do not experience this problem, its memory voltage by default is 1.6v and also uses Hynix memory.

2. Yes I monitor vRAM usage with GPU-Z 0.6.7 - Memory Usage (Deadicated) sensor reading.

3. Attached is a copy of the modified version of Trixx 4.4.0b. This version allows for adjustment of memory voltage, the standard version of Trixx 4.4.0b does not. Also included is a picture showing the modded version of Trixx with memory voltage set to 1.6v, also with GPU-Z verifying the voltage is actually set and also showing the card with a vRAM load over 1600MB running fine (GPU load 99%).

4. My full system spec:

intel Core i5 2500k
Asrock Z77 Extreme6 w\ UEFI 2.6
2x4GB G.Skill Ripjaws PC12800 7-8-7-24
640GB Western Digital Caviar Black, SATA2 32MB Cache
2x SATA Sony Optiarc DVD-RW
Asus Xonar DX 7.1 (unixonar drivers)
Corsair HX 850w
Windows 7 Ultimate SP1
All latest drivers for chipset, graphics, sound, etc

5. I have tried setting 1.55v for the memory but stability is patchy, not as bad, but still not 100% stable. The absolute minimum memory voltage I can set that is seemingly stable is 1.58v, but I set memory voltage to 1.6v to factor in such things as the possibility of memory vdroop. Both Elpida and Hynix (Vapor-X I have uses Hynix memory) verify that at higher memory frequencies such as what the Vapor-X runs at memory voltage needs to be in the region of 1.55v - 1.6v, depending on the application of the memory and choice of components for the design. You can see this for yourself here;

Elpida; http://www.elpida.com/en/products/gddr5.html
Hynix; https://www.skhynix.com/products/gra...No=H5GQ2H24MFR

I have also tested GPU vdroop levels, when GPU vdroop levels were tested Heaven 3.0 was used with the following settings (Heaven doesn't use as much memory as the other tests so the card doesn't crash):

Tessellation: Extreme
Shaders: High
AA: 8x
AF: 16x
Resolution: 1680*1050

Board Li power was set to +20 then measurements of GPU vdroop were taken with GPU-Z 0.6.7 while Heaven 3.0 was left to loop. I also let a colleague at work measure GPU vdroop directly via hardware as he works more hands-on with the actual hardware side of things than I do. His findings were similar to what GPU-Z reported, vdroop is anywhere in the region of 0.025v to 0.078v depending on the level of GPU usage and the level of AA / AF. With a 100% GPU utilisation, Extreme Tessellation and 8x AA, 16x AF applied GPU vdroop was as bad as 0.078v and never better than 0.050v.

I've also had to completely redesign the VRM cooling on the Vapor-X, at stock frequencies with Heaven 3.0 using the above settings VRM temps hit in excess of 126c very rapidly, suffice to say and unsurprisingly the card first started having wild GPU frequency fluctuations (constantly bouncing between 501MHz and 950MHz) then soon after the card crashed with VRM temps that high. During this period GPU temps were never hotter than 65c (measured with a infrared temp gun thats accurate to +/- 2%)

Regards, Dave

---

I got one of these cards and I am having second thoughts about keeping it. I was hoping Sapphire would have fixed the BIOS by now and the delay of the newest TRIXX software that is supposed to help over clock this card has got me worried.


I think the main problem with this card is the terrible VRM cooling and i don't like the RAM making contact with the heat sink. I think RAM is better cooled by itself. Adding the heat to the same heat sink that is cooling the GPU is only going to sponge up extra heat and the ram seems gets hot to the touch on my other HD7950.


I think Sapphire is holding back on releasing TRIXX or BIOS updates because they know the poor heat sink design isn't good enough to support higher voltage and speeds without over heating.

I think this card with a well made proper fitting water block and good BIOS and TRIXX software would over clock very well. Too bad nobody makes a proper full cover water block for it.

Do you guys know if the Elpida memory is just as good as the HYNX?

To me it seems like they should be rated for the same speeds and be similar to performance. For over clocking I would like the best. I'm disappointed that there is no consistency on the parts and design of this card. Seems like Sapphire is just using up parts available and sticking the VAPOR X name on it to sell it fast.

Elpeda memory should OC just as well as Hynix, in both cases the memory will need 1.6v, otherwise the card may be unstable even at stock frequencies never mind when OCing. There are plenty of people saying their card is stable with 1.6v memory voltage but not 1.5v. Really the only hope you have is if you get a card with Hynix memory as theres a vBIOS floating around for these cards giving the memory 1.6v. Sapphire are fast running out of time with me, they have been just about as unhelpful as they possibly can be - stalling, inability to answer the simplest of questions, currently no reply to my last email response categorically pointing the finger at low memory voltage, absolutely no word on the release of Trixx 4.6 even though some reviewers have it, the list goes on and on and on. Worse still, I've done their job for them and they STILL can't get their thumb out of their arses. Unless Sapphire at least manage to release a vBIOS for these cards increasing memory voltage and make Trixx 4.6 available, I intend to spread the word even further to hurt their reputation as much as possible. Issues like this should be dealt with properly, not quietly shoved into the long grass. I'd go as far to say as that anyone in the US has some solid ground for a lawsuit against Sapphire.

Elpida memory sucks hard for overclocking,stick with hynix ;)

I've had Elpeda memory on some cards, it's OCd decently enough managing to reach ballpark average figures from those of Hynix and Samsung. The only warning I would give about Elpeda memory is that all the cards I've had that used it mysteriously died about 15 months later :hm:

Hi...can I fix on this card reference waterblock 7970? thanks

Quote:

---

Originally Posted by jaaa1976

Hi...can I fix on this card reference waterblock 7970? thanks

---

No.


PCB of this card is so unique that no full cover water block will fit it. A Swiftech MCW82-7900 will fit it but it will only cover the GPU.



I had problems with voltage control on my first Vapor X HD7950 and I got rid of it and ordered another and the new one does do voltage control but only when booting into the BIOS that is selected with the button in the non illuminated position.


I haven't experienced poor VRM heat problems like Ket has. My VRMs don't heat up very much at all. Maybe the VRM heat sink was not tight enough.

Another interesting thing that I found is that my Elpida memory doesn't over clock well at all. I can get stable at around 1350mhz on the memory with the memory voltage at default setting of 1.6v. However I can under volt the memory to 1.45v and the memory can operate higher at 1400mhz.

I really think the heat sink design is hurting the cards ability to over clock for the following reason.

1. The heat of the the ram and GPU in the heat sink act like a sponge and hurt each other. Better design for cooling would be to keep memory heat away from the GPU. I think the only reason SApphire decided to make contact with the heat sink is more to support the weight of the heat sink rather than to aid the memory's cooling.


2. I think the memory voltage at 1.6v may be too high because it over heats the ram too fast causing it to be unstable and hurts GPU ability to cool down. The problem of the high voltage on the ram mixed with the added heat of the gpu is hurting the performance of the ram and GPU. Elpida designed this memory to run at a set speed with a set voltage without any type of heat sink on it...open air. When the ram doesn't have good air flow and makes contact with a hot surface of the block I'm sure the ram's electrical performance goes down the drain.

Anybody found a vBIOS for this card newer than 15.25.99? I'll get back to this issue when I've got a little more time. I haven't forgotten about it ;)

Quote:

---

Originally Posted by Ket

Anybody found a vBIOS for this card newer than 15.25.99? I'll get back to this issue when I've got a little more time. I haven't forgotten about it ;)

---

Mine is 015.025.000.099.000000 (113-210PBBBG-V07) as reported by GPU-Z.

I really think the problem with this card for over clocking is the BIOS.

My Vapor X HD7950 over clocked after using a BIOS editor on the BIOS.

http://minidriven.com/WARVAPORXHD7950/OCVX7950.jpg

I have water cooled it and I think it has more room on core for over clocking but I just have to figure out how to tweak voltage.

http://minidriven.com/WARVAPORXHD795...XHD7950WAR.jpg


The Elpida memory is tough to over clock. 1500mhz is about as fast as I can get it to go with 1.55v.


The card games great.

Thats a real shame about your card Sonda, I'm still looking at fixing this broken POS when I have time. One thing I can say for certain is even after all the VRM cooling mods I've done (got them down to a max of around 90c @ 1150 / 1600 1.056v (actual), 1.6v memory running unigine valley maxed) is that these VRMs do not like any kind of voltage at all. Even with a paltry 1.056v (0.98v under extreme load without me compensating for vdroop) the VRMs still hit 80c+, that kind of temp for voltage that low and horrible vdroop is just a joke. I could throw almost 1.2v at my 7950 TwinFrozr with its stock cooler with the same GPU speed and faster memory speed and the VRMs wouldn't get hotter than about 83c. If I had gone to the lengths with that card as I have with this Vapor-X I doubt the VRMs would get much hotter than 70c or so.

Not sure what to do with this card, on one hand by keeping it I'll continue to push forward trying to make a nice vBIOS editor, on the other hand I'm sick to death of this Vapor-X and just want to swap it for a GTX670 or something. Hell, at this point I'd even consider swapping it for a excellent clocking HD6970 as far as I recall those cards didn't suffer horrid VRM temps once decent VRM heatsinks were fitted.