• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Time for a Purchase (HD 7970)

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I don't possess the in-depth knowledge that many of the others posters here have, but I do own a 7970, this one, to be exact, and I am thrilled with its performance. As I type this, it is hashing away at 1135MHz, 1109mV actual, and 66°C at 21°C ambient. It can go higher, but I like to keep the temp down and fan speed below 70% (they are a little noisy when spooled up). Some people do get incredible results with the 7950, it's true. I guess what you get when you buy a good 7970 is ease of getting there, I believe with the 7950 it takes a bit more effort (special cooling, too?) to equal or exceed a good 7970 on air.
 
I don't think anyone is trying to say the 7970 isn't faster than the 7950. Just not enough faster (with both O/C'd) to make a noticeable difference while gaming. There's the $100 price difference, as well to consider.
 
3DVagabond said:
I don't think anyone is trying to say the 7970 isn't faster than the 7950. Just not enough faster (with both O/C'd) to make a noticeable difference while gaming. There's the $100 price difference, as well to consider.
Click to expand...

I'll stipulate that the 7970 is above the "knee" of the price/performance curve. I should have confined my comments to the perceived quality of the voltage unlocked Sapphire unit v. the more questionable (in my eyes) XFX unit.
 
crashtech said:
I don't possess the in-depth knowledge that many of the others posters here have, but I do own a 7970, this one, to be exact, and I am thrilled with its performance. As I type this, it is hashing away at 1135MHz, 1109mV actual, and 66°C at 21°C ambient. It can go higher, but I like to keep the temp down and fan speed below 70% (they are a little noisy when spooled up). Some people do get incredible results with the 7950, it's true. I guess what you get when you buy a good 7970 is ease of getting there, I believe with the 7950 it takes a bit more effort (special cooling, too?) to equal or exceed a good 7970 on air.
Click to expand...

The question becomes, can I get a CPU/Mobo/Ram/SSD/Case for 600 once I get the 7970 haha. I think I'll exclude the Case from the price since I'll reuse it for every build to come and it's the only thing that has to be amazing. I bought a shoddy case once, and I will never get one again. The worst thing to have is a terrible environment to work in that is loud and obnoxious with stupid LEDs. I have a 1k soft limit.

Reason I wanted the 7970 is I just wanted something semi top of the line for once haha. A little future proofing also, people are still playing new games on their top of the line 6970s and other cards with ultra settings. That and I'm hoping that CF is fixed soon so I can one day consider it although that's something I know is a strech considering thermals, mobo required, etc.

Like I've said though, I am currently playing Dark Souls at 1080p with DSFix on a 9800M GTS on my laptop and I think the graphics are really good haha. Part of me wants to be blown away and realize how much better life can be! I'm trying to replicate the console experience with a Desktop though, and the test drive using my gaming laptop is going great so I can't wait no matter what card I pick. I just have a 1k soft budget to work with. I think I'm owed over 600 dollars though, second I find my other phone I'll know, so I have more money coming in to work with.
 
tential said:
The question becomes, can I get a CPU/Mobo/Ram/SSD/Case for 600 once I get the 7970 haha. I think I'll exclude the Case from the price since I'll reuse it for every build to come and it's the only thing that has to be amazing. I bought a shoddy case once, and I will never get one again. The worst thing to have is a terrible environment to work in that is loud and obnoxious with stupid LEDs. I have a 1k soft limit.

Reason I wanted the 7970 is I just wanted something semi top of the line for once haha. A little future proofing also, people are still playing new games on their top of the line 6970s and other cards with ultra settings. That and I'm hoping that CF is fixed soon so I can one day consider it although that's something I know is a strech considering thermals, mobo required, etc.

Like I've said though, I am currently playing Dark Souls at 1080p with DSFix on a 9800M GTS on my laptop and I think the graphics are really good haha. Part of me wants to be blown away and realize how much better life can be! I'm trying to replicate the console experience with a Desktop though, and the test drive using my gaming laptop is going great so I can't wait no matter what card I pick. I just have a 1k soft budget to work with. I think I'm owed over 600 dollars though, second I find my other phone I'll know, so I have more money coming in to work with.
Click to expand...

Just get the Sapphire 7950 and pocket the $80-100 you'll save. Clock for clock, a 7950 is only ~5% slower than a 7970. http://hexus.net/tech/reviews/graphics/34761-amd-hd-7950-vs-hd-7970-clocks/

All of the 7950s on the market today come with Boost BIOS which automatically overclocks to 925MHz if you have the thermal headroom for it (crank up Powertune to +20% to make sure you do). 925MHz is the stock clock for regular 7970s.

So the 7950s available today with Boost BIOS are within 5% of stock 7970s.

If you really want to spend $100-150 more for 5-10% more performance with the 7970s or 7970 GHz editions, go for it, but don't think you're going to futureproof by spending an extra for a lousy 10% more performance... that's just laughable. 10% is like the difference between 40 fps and 44 fps... most people can't tell, period. You're just throwing money away.

Bottom line: get that Sapphire 7950 and use the money you save on something else.
 
If you are planning on getting a 7950, this is the best 7950 at MSRP or less

www.newegg.com/Product/Product.aspx?...82E16814202030

SAPPHIRE 100352-4L Radeon HD 7950 3GB GDDR5 PCI-Express Video Card with Boost

It's using 7970 pcb with high ASIC quality chips.

Spending more than 300 dollars on a 7950 is truely a rip off, just like spending more than 400 dollars on a 7970 is a rip off.

If you don't plan on overclocking ever,

http://www.newegg.com/Product/Produc...82E16814161420

HIS IceQ Boost Clock H795QC3G2M Radeon HD 7950 3GB 384-bit GDDR5 PCI Express 3.0 x16 HDCP Ready CrossFireX Support Video Card

It has inferior VRMs but has an absolutely amazing blower, which is important if you plan on going crossfire in the future.
 
Communism said:
If you are planning on getting a 7950, this is the best 7950 at MSRP or less

www.newegg.com/Product/Product.aspx?...82E16814202030

SAPPHIRE 100352-4L Radeon HD 7950 3GB GDDR5 PCI-Express Video Card with Boost

It's using 7970 pcb with high ASIC quality chips.

Spending more than 300 dollars on a 7950 is truely a rip off, just like spending more than 400 dollars on a 7970 is a rip off.

If you don't plan on overclocking ever,

http://www.newegg.com/Product/Produc...82E16814161420

HIS IceQ Boost Clock H795QC3G2M Radeon HD 7950 3GB 384-bit GDDR5 PCI Express 3.0 x16 HDCP Ready CrossFireX Support Video Card

It has inferior VRMs but has an absolutely amazing blower, which is important if you plan on going crossfire in the future.
Click to expand...

First off, those Sapphire "reference" 7950s are NOT using 7970 reference PCBs like you claimed! The plastic parts may be the same but the power circuitry is NOT like a reference 7970 PCB! These wannabe-reference-7970 PCBs are missing a power phase, and the connectors are 6+6 and not 8+6. So it's not true to say that they use higher-quality 7970 reference PCBs.

Second, I have had three of those semi-reference 7950s, and their ASICs were all below 70% which is not high, so no, they don't bin chips or anything, it's still a lottery as far as what you get. And ASIC doesn't mean much by itself. One of them just could not stay stable, it was HORRIBLE. At an ASIC of 68% you should not have to be like 1.2 volts just to keep stock clocks stable, but that was where I wound up because it was such a crappy card.

This Sapphire Dual-X 7950 runs cooler and quieter for only $10 more ($290 after rebate): http://www.newegg.com/Product/Produc...82E16814202026

It really depends on your priorities and usage scenarios as to which one is the best. Since he's not going crossfire, and by the time he needs crossfire he might as well just sell his old card and get a HD 9950 or whatever, I don't think he can go far wrong either way but the Dual-X will be quieter for sure. (I have a Sapphire Dual-X 7970 as well as a Sapphire Dual-X 7950, by the way.)

Furthermore that HIS card you linked to reported is a 2-slot card that has a cooler so big that it actually intrudes on the slot below that one, so it actually needs 3 slots, so you can't crossfire them in motherboards where you don't get two full slots between PCIe 16x/8x slots. Even if you do have two full slots like with most modern high-end motherboards, it doesn't sound like it'd be so good for crossfiring anyway, as fans lose efficiency when you don't let them breathe.
 
Last edited:
blastingcap said:
I have three of those semi-reference 7950s, and their ASICs were all below 70% which is not high, so no, they don't bin chips or anything, it's still a lottery as far as what you get. Furthermore, they are NOT using 7970 reference PCBs like you claimed!!! The size is the same but they are missing a power phase, and the connectors are 6+6 and not 8+6. So it's not true to say that they use higher-quality 7970 reference PCBs... I mean technically the plastic parts of the PCB are the same I guess, but the power circuitry is inferior.

This Sapphire Dual-X 7950 runs cooler and quieter for only $10 more ($290 after rebate): http://www.newegg.com/Product/Produc...82E16814202026

It really depends on your priorities and usage scenarios as to which one is the best. Since he's not going crossfire, and by the time he needs crossfire he might as well just sell his old card and get a HD 9950 or whatever, I don't think he can go far wrong either way but the Dual-X will be quieter for sure. (I have a Sapphire Dual-X 7970 as well as a Sapphire Dual-X 7950, by the way.)
Click to expand...

Haven't seen anyone with a Sapphire 4L with lower than 60% asic, while the majority of Sapphire 2L and 3L come at 40-60% asic quality, so yes, its a big difference.

Also, Reference 7950 still is worse than the Sapphire 4L, so it is superior to Reference 7950.

MSI 7950 and HIS 7950 and XFX 7950 currently are 40-60% asic quality.

The difference between 50% asic quality and 70% asic quality is massive in TDP terms.

3Ls are also inferior to 4Ls and 2Ls, making it not worth more than the 4L (you can't get 2Ls for 300 dollars or less anymore, or I would have recommended it)
 
Communism said:
Haven't seen anyone with a Sapphire 4L with lower than 60% asic, while the majority of Sapphire 2L and 3L come at 40-60% asic quality, so yes, its a big difference.

Also, Reference 7950 still is worse than the Sapphire 4L, so it is superior to Reference 7950.

MSI 7950 and HIS 7950 and XFX 7950 currently are 40-60% asic quality.

The difference between 50% asic quality and 70% asic quality is massive in TDP terms.

3Ls are also inferior to 4Ls and 2Ls, making it not worth more than the 4L (you can't get 2Ls for 300 dollars or less anymore, or I would have recommended it)
Click to expand...

Please stop thinking that high ASIC means anything across models. It doesn't. A 71% of one model might be better than a 61% of the same model, but might be WORSE than a 51% of yet another model. And most companies have multiple models.

I have handled two different MSI TFIII's at 89% and 91% ASIC, as well as a 71% and 61% ASIC. Guess what, the TFIIIs ate more power and ran hotter and louder than the Sapphire Dual-X at 61% ASIC and drew 40 more watts. Same loads, clocks, and both were non-Boost. I did some further tests and undervolted the MSI to a more extreme degree than the Dual-X to see if that could close the gap, but regardless of whether it was equal voltage or lower voltage the result was the same: the TFIII ate more power to produce the same performace. And that was true even after I replaced the crappy thermal paste job on the MSI TFIII's and undervolted the MSI TFIII way more than I undervolted the Sapphire Dual-X. Those dinky little fans didn't help, and I suspect the power circuitry was inferior as well. I guess the fans did such a crappy job of cooling that the GPU heated up and that temperature rise erased whatever advantage it should have had via a lower voltage. Whatever the actual reason, the Sapphire Dual-X with 61% ASIC score kicked the ass of the MSI TFIII with 91% ASIC. It wasn't even close.

Another example: I recently handled a MSI 7790 and a ASUS 7790 DirectCU II which has a visually superior dual-fan, huge copper heatpipes cooler compared to the single big fan, skinny heatpipes MSI cooler. The ASUS had a 81% ASIC. The MSI had a 73% ASIC. Now you would expect the card with higher ASIC GPU and much larger cooler with larger heatpipes and two fans to be lower wattage and run cooler, right? In theory, maybe. But in practice, the ASUS ate 10 more watts at the wall than the MSI according to my KillaWatt monitor and actually ran slightly hotter. Identical load, identical clocks, and voltage was very comparable and may have even favored the ASUS. (Voltage control not available for 7790s yet, even the DirectCUIIs, so I looked at the GPU-Z readouts as they rose and fell). Once again, the card with the bigger fan and lower ASIC actually drew LESS power and produced LESS heat than the card with smaller fans and higher ASIC.

So please stop spreading misinformation that ASIC means anything across models. It doesn't. Although I would agree that within the same model, higher ASIC might help.

Also, your claim about how MSI hasn't made 7950 cards with ASICs above 60% lately is blatantly wrong. I handled a recent vintage MSI TFIII Boost 7950 with 71% ASIC and another was 61% ASIC.

A word about 7950 reference PCBs: there were two reference 7950s (one had axial, one had radial fan) so your claim that the 4L PCB is better than reference is inaccurate if we're comparing blower vs blower. It apparently *is* the 7950 reference PCB (blower type). But both are inferior to the 7970 PCB which is what you initially claimed it was. It is definitely NOT the 7970 PCB. It's missing a power phase and is a 6+6 board, NOT 8+6.

I can't speak to the HIS and XFX ASICs but like I stated above, you can't really compare ASIC scores between different models as there is more to power and thermals than ASIC. I learned that lesson with the MSI TFIII (see above).

TL;DR version of what I said: ASIC scores should not be compared across models as they can be misleading. I don't know exactly why, but I suspect bigger fans and better PCB circuitry can more than offset ASIC score advantage, so it's dangerous to compare ASIC scores between two different models.
 
Last edited:
blastingcap said:
Please stop thinking that high ASIC means anything across models. It doesn't. A 71% of one model might be better than a 61% of the same model, but might be WORSE than a 51% of yet another model. And most companies have multiple models.

I have handled two different MSI TFIII's at 89% and 91% ASIC, as well as a 71% and 61% ASIC. Guess what, the TFIIIs ate more power and ran hotter and louder than the Sapphire Dual-X at 61% ASIC and drew 40 more watts. And that was true even after I replaced the crappy thermal paste job on the MSI TFIII's and undervolted the MSI TFIII way more than I undervolted the Sapphire Dual-X. Those dinky little fans didn't help, and I suspect the power circuitry was inferior as well. I guess the fans did such a crappy job of cooling that the GPU heated up and that temperature rise erased whatever advantage it should have had via a lower voltage. Lesson learned: avoid video cards with dinky little fans regardless of how binned the chips are or how impressive the build quality appears to be.

Another example: I recently handled a MSI 7790 and a ASUS 7790 DirectCU II which has a visually superior dual-fan, huge copper heatpipes cooler compared to the single big fan, skinny heatpipes MSI cooler. The ASUS also had a 81% ASIC. The MSI had a 73% ASIC. Yet the ASUS ate 10 more watts at the wall than the MSI according to my KillaWatt monitor. Same load, same everything.

So please stop spreading misinformation that ASIC means anything across models. It doesn't. And also, your info about ASICs is wrong, wrong, wrong. I handled a recent vintage MSI TFIII Boost 7950 with 71% ASIC and 61% ASIC which blows apart your claim that the non-4L 7950 ASICs are all less than 60%.

There were two reference 7950s (one had axial, one had radial fan) so your claim that the 4L PCB is better than reference is inaccurate if we're comparing blower vs blower. It apparently *is* the 7950 reference PCB (blower type). But both are inferior to the 7970 PCB which is what you claimed it was. It is definitely NOT the 7970 PCB as it's missing a power phase and is a 6+6 board, NOT 8+6.

I can't speak to the HIS and XFX ASICs but like I stated above, you can't really compare ASIC scores between different models as there is more to power and thermals than ASIC. I learned that lesson with the MSI TFIII which is lauded for being high-ASIC but in practice I have found it to be no better, and possibly worse, than Sapphire's Dual-X cards. Even with a stronger UNDERvolt than the Sapphire Dual-X, at same clocks the Sapphire Dual-X ASIC 61% drew 40 watts LESS. This was true before AND after I replaced the TFIII's thermal paste.
Click to expand...

I made it clear I was talking about cards that are 300 dollars or less.
Don't bring up cards that shot up to 320 or 330 dollars because of their superior design and therefore destroy the value proposition of the 7950 (the whole reason you would get one in the first place)

Clock speed, ASIC and voltage determine TDP when using same power circuitry. Card temperature != TDP.

Reference 7950 PCB is 2+1 Power phase for 3 Phase Total.
Reference 7970 PCB is 5+1 Power phase for 6 Phase Total.

The Reference 7970 PCB used on the Sapphire 4L and the HIS IceQ X^2 is 4+1 Power phase for 5 Phase Total.

The reason people have this confused is because the review units that went to the reviewers were all using the Reference 7970 PCB with 4+1 Power phase for 5 Phase Total since the Reference 7950 PCB was not ready for production at the 7950 Launch.

EDIT:
Whatever, change your post and keep spreading you FUD all you want. You only change you tone momentarily when you're called out on it and then change it right back the very next post. I'm done talking to you after having to read your FUD posts on 14+ threads now.

EDIT2: Public Service Announcement: blastingcap is a psychopath and therefore is poisonous to your mind. His posts contain designed logical flaws to fuk with your brain deliberately and to leave the game to be played with more traps. He's changed his argument at least 7 times in just this thread. If you want to see an entertaining thread read the bitcoin one, where he changes his argument diametrically back and forth every other post for about 10 pages. I don't know why psychopaths are not murdered as children anymore like was the evolutionarily biological solution for this in the past.
Guess I'm just too trusting of people since I'm a communist.

Personal attacks are not permitted. This post clearly steps well over any line. -Admin DrPizza
 
Last edited by a moderator:
What? The Sapphire Dual-X 7950 is selling for $290 after rebate, and furthermore the HIS is also in the $300 ballpark as well, sans rebate. The -4L is $280 after rebate right now. So we ARE talking about sub-$300 cards.

Re: "Card temperature != TDP." Well yeah, duh. But I used a Killawatt to measure wattage so I am not talking just thermals. The higher ASIC cards ate more wattage than the lower ASIC cards of different models.

I have a lot of experience with playing with voltage and clocks and measure it all with a Killawatt that I have no reason to believe is defective. I did like a triple take the first time I saw the results of my TFIII test. I thought maybe something was wrong due to thermal paste since MSI thermal paste jobs are horrible, but even after I cleaned their mess up and used MX-4 (a top thermal paste very cloes to PK-1 and Gelid GC3), I got the same stunning result. Still not convinced, I tested the ASIC theory with the ASUS and MSI 7790s. I still don't know why I got those results, but the only thing I can say for sure is that one should not compare ASIC scores of two different models of cards, because there is apparently a lot more than goes into wattage than ASIC score. One thing is for sure, though: I am going to avoid small-fan video cards from now on, just in case that's a major factor (which is probably is... all fans are NOT created equal and the MSI TFIII fans don't look like they produce nearly as much static pressure or airflow as the 32% larger, scalloped Sapphire Dual-X fans).

For the second time, even if we take everything you say as true, the point is that the -4L 7950s do NOT use the full 7970 PCB with all power phases and 8+6 connectors which is what you implied when you said it uses the 7970 PCB. That (comparing it to reference 7970) is the issue, not comparing it to -2L or -3L or reference or whatever else. I think we both agree that the -4L PCB is outclassed by the 7970's PCB. It's apparently tied with the original reference 7950's PCB, which is a notch below. That is all. I think we can agree to this and can drop this topic now.

Communism said:
I made it clear I was talking about cards that are 300 dollars or less.
Don't bring up cards that shot up to 320 or 330 dollars because of their superior design and therefore destroy the value proposition of the 7950 (the whole reason you would get one in the first place)

Clock speed, ASIC and voltage determine TDP when using same power circuitry. Card temperature != TDP.

Reference 7950 PCB is 2+1 Power phase for 3 Phase Total.
Reference 7970 PCB is 5+1 Power phase for 6 Phase Total.

The Reference 7970 PCB used on the Sapphire 4L and the HIS IceQ X^2 is 4+1 Power phase for 5 Phase Total.

The reason people have this confused is because the review units that went to the reviewers were all using the Reference 7970 PCB with 4+1 Power phase for 5 Phase Total since the Reference 7950 PCB was not ready for production at the 7950 Launch.

EDIT:
Whatever, change your post and keep spreading you FUD all you want. You only change you tone momentarily when you're called out on it and then change it right back the very next post. I'm done talking to you after having to read your FUD posts on 14+ threads now.
Click to expand...

EDIT: What FUD? You are the one spreading FUD with your bogus claims:

- reference 7970 PCB? WRONG given that you didn't give the caveat about the missing power phase. (NOTE: It occurs to me that maybe why my MSI TFIII had higher power draw was because of it had a different number of power phases than the Sapphire. I don't know if it did or not, I am just hypothesizing. I'll have to look into that some more.)
- MSI doesn't make 7950s with ASIC higher than 60% anymore? WRONG.
- higher ASIC necessarily means lower wattage? I can't say for sure if this is always false, but it's definitely not always true. I've seen it with my own eyes that you can't compare ASIC between models. And I had a 68% ASIC 7950 that couldn't overclock to save its life. Maybe it was a rare defective one or something, but I saw it with my own eyes.

I edited my post only to make it clear I wasn't talking just about thermals but wattage as well, for both the TFIII and ASUS, and to clean up the paragraphs but the fundamental post is the same. I didn't do any "180" that I know of. Please educate me if I did. And unless my Killawatt broke, I know what I saw with my own eyes. It's not like I edited to say anything nefarious after the fact or whatever.

If you want to be reasonable go ahead and tell me what I stated was wrong.

Else I think you're just being defensive because you stated three wrong things in this thread alone, and apparently can't stand to be called out on it. I have no idea what other 13 threads you are talking about or why you are so OCD as to actually count how many there are. Quit stalking me, bro.
 
Last edited:
OP if you want good value for money and a well built card go for HIS HD 7950 Iceq x2 boost. built on a reference HD 7970 PCB. 8 + 6 pin power. more power phases than reference HD 7950. unlocked voltage. excellent cooler. good VRM cooling. probably the best built HD 7950.

check if you can fit the card in your case as its 311 mm long. at the same clocks HD 7950 is 5% slower than HD 7970. average overclocks for this card are 1150 -1200 mhz.

http://www.newegg.com/Product/Produc...82E16814161429

http://www.hisdigital.com/un/product2-731.shtml

http://www.legionhardware.com/artic...hz_edition_7950_iceq_xsup2_boost_clock,3.html
http://www.techpowerup.com/reviews/HIS/HD_7950_X2_Boost/31.html
http://www.overclockers.com/his-hd-7950-iceq-x2-boost-clock-graphics-card-review
 
raghu78 said:
OP if you want good value for money and a well built card go for HIS HD 7950 Iceq x2 boost. built on a reference HD 7970 PCB. 8 + 6 pin power. more power phases than reference HD 7950. unlocked voltage. excellent cooler. good VRM cooling. probably the best built HD 7950.

check if you can fit the card in your case as its 311 mm long. at the same clocks HD 7950 is 5% slower than HD 7970. average overclocks for this card are 1150 -1200 mhz.

http://www.newegg.com/Product/Produc...82E16814161429

http://www.hisdigital.com/un/product2-731.shtml

http://www.legionhardware.com/artic...hz_edition_7950_iceq_xsup2_boost_clock,3.html
http://www.techpowerup.com/reviews/HIS/HD_7950_X2_Boost/31.html
http://www.overclockers.com/his-hd-7950-iceq-x2-boost-clock-graphics-card-review
Click to expand...

TechPowerup said it's 8+8 pin (http://www.techpowerup.com/reviews/HIS/HD_7950_X2_Boost/4.html), but Newegg says 8+6 pin, hmmmmm. OP should also make sure his case can take something like 12 inches, because that's about how long the card is, according to Newegg reviews. But yes, nice card. HIS has come a long way with the 7xxx series coolers, compared to their mediocre 6xxx coolers.
 
blastingcap said:
TechPowerup said it's 8+8 pin (http://www.techpowerup.com/reviews/HIS/HD_7950_X2_Boost/4.html), but Newegg says 8+6 pin, hmmmmm. OP should also make sure his case can take something like 12 inches, because that's about how long the card is, according to Newegg reviews. But yes, nice card. HIS has come a long way with the 7xxx series coolers, compared to their mediocre 6xxx coolers.
Click to expand...

I remember when that review was released. I've never seen the 8+8 7950 in that review. The 7970 is 8+8. Maybe W1zzard mixed up some pics? Or, there was a mix up with HIS on the review sample? I dunno?
 
3DVagabond said:
I remember when that review was released. I've never seen the 8+8 7950 in that review. The 7970 is 8+8. Maybe W1zzard mixed up some pics? Or, there was a mix up with HIS on the review sample? I dunno?
Click to expand...

You're probably right. He probably mixed up the pics, because I just did a little more searching, and everyone else says 8+6, some with photos clearly showing 8+6.
 
blastingcap said:
Please stop thinking that high ASIC means anything across models. It doesn't. A 71% of one model might be better than a 61% of the same model, but might be WORSE than a 51% of yet another model. And most companies have multiple models.

I have handled two different MSI TFIII's at 89% and 91% ASIC, as well as a 71% and 61% ASIC. Guess what, the TFIIIs ate more power and ran hotter and louder than the Sapphire Dual-X at 61% ASIC and drew 40 more watts. Same loads, clocks, and both were non-Boost. I did some further tests and undervolted the MSI to a more extreme degree than the Dual-X to see if that could close the gap, but regardless of whether it was equal voltage or lower voltage the result was the same: the TFIII ate more power to produce the same performace. And that was true even after I replaced the crappy thermal paste job on the MSI TFIII's and undervolted the MSI TFIII way more than I undervolted the Sapphire Dual-X. Those dinky little fans didn't help, and I suspect the power circuitry was inferior as well. I guess the fans did such a crappy job of cooling that the GPU heated up and that temperature rise erased whatever advantage it should have had via a lower voltage. Whatever the actual reason, the Sapphire Dual-X with 61% ASIC score kicked the ass of the MSI TFIII with 91% ASIC. It wasn't even close.

Another example: I recently handled a MSI 7790 and a ASUS 7790 DirectCU II which has a visually superior dual-fan, huge copper heatpipes cooler compared to the single big fan, skinny heatpipes MSI cooler. The ASUS had a 81% ASIC. The MSI had a 73% ASIC. Now you would expect the card with higher ASIC GPU and much larger cooler with larger heatpipes and two fans to be lower wattage and run cooler, right? In theory, maybe. But in practice, the ASUS ate 10 more watts at the wall than the MSI according to my KillaWatt monitor and actually ran slightly hotter. Identical load, identical clocks, and voltage was very comparable and may have even favored the ASUS. (Voltage control not available for 7790s yet, even the DirectCUIIs, so I looked at the GPU-Z readouts as they rose and fell). Once again, the card with the bigger fan and lower ASIC actually drew LESS power and produced LESS heat than the card with smaller fans and higher ASIC.

So please stop spreading misinformation that ASIC means anything across models. It doesn't. Although I would agree that within the same model, higher ASIC might help.

Also, your claim about how MSI hasn't made 7950 cards with ASICs above 60% lately is blatantly wrong. I handled a recent vintage MSI TFIII Boost 7950 with 71% ASIC and another was 61% ASIC.

A word about 7950 reference PCBs: there were two reference 7950s (one had axial, one had radial fan) so your claim that the 4L PCB is better than reference is inaccurate if we're comparing blower vs blower. It apparently *is* the 7950 reference PCB (blower type). But both are inferior to the 7970 PCB which is what you initially claimed it was. It is definitely NOT the 7970 PCB. It's missing a power phase and is a 6+6 board, NOT 8+6.

I can't speak to the HIS and XFX ASICs but like I stated above, you can't really compare ASIC scores between different models as there is more to power and thermals than ASIC. I learned that lesson with the MSI TFIII (see above).

TL;DR version of what I said: ASIC scores should not be compared across models as they can be misleading. I don't know exactly why, but I suspect bigger fans and better PCB circuitry can more than offset ASIC score advantage, so it's dangerous to compare ASIC scores between two different models.
Click to expand...

Very interesting read there, I was kinda disappointed in my GB WF3 7950 because of its 67% ASIC score but that seems to be even slightly above average I was also under the impression it's a 7970 pcb since it's 8+6 pin design. I still wouldn't recommend it since its volt locked but it does boost to 1ghz stock without any issues.
 
Never heard of ASIC unsurprisingly since I haven't followed GPU advancement for 5 years. Interesting.

I do like the 7950 idea. However, even if clock for clock it's only 5% slower, why would I clock them the same? I never would unless I'm missing something.

I guess the main argument is what am I missing if I get a 7950 vs a 7970? Any games that I couldn't run on 1080p maximum settings? CF was only out of the question because I don't really believe I could afford a mobo like that. But mobos are the thing I understand least on a PC.

The problem with reading reviews for me is that every review is testing the cards to their maximum potential to show differences. I may not need to use 4x MSAA sometimes or notice the difference compared to FXAA. With no accompanying pictures, I really have no clue what the actual difference is when a card can't run a particular setting. Seeing some NVidia guides has helped recently in highlighting some of the differences in certain games, but some of that to me is just marketing to sell their higher end GPUs.

Biggest reason I started looking at the 7970 was that I figured, I can run everything now, but 2 years down the road, if I'm not able to run it, or I just want more, I figure it'd be easier to sell a 7970, and get something else. And I could recoup some of it with mining too, second I figure that out haha. So I could own a high end desktop card (my first high end desktop card), and be able to get quite a large amount back(150-250 back).
-------
The plan currently though is:
CPU: Intel Core i7-3770K or Intel Core i5-3570K (Re-encoding all the videos to my server so that they're smaller is going to most likely have to be handled by this machine so I thought a 6 core CPU might help with that.)
Mobo: No clue but 150 dollar limit on it. I really never know what to look for.
Ram: 8 GB of ram.
GPU: 7970 LE was what I was looking at based off of Tomshardware's love of that card at that 250-300 price range. 7970 hit me though because it's dipping below 400.
Case: 200-300 dollar limit seperate budget. Silence, Aesthetics that's what I'm doing with this. It's going to be in the open in a living room type setting. My current case is a NZXT case that screams gamer nerd with LEDs that I hate.
HD: Small Solid State. Most of my information is stored on my old PC that acts like a server. Some weird retailer is selling old servers for 360-400 with server cases that would cost 200 by themselves so all my hard drives/storage will be in there. My PC only needs to hold the 2-3 games currently installed. I have 6+ terabytes of storage already, and will have 16 TB within 2 years(probably much more at the rate I'm going with movies/tvshows).
 
Hey OP the Double D cooler is a dirty beast and you don't want it. Mine couldn't even cool a 7870 properly. ASUS and Gigabyte seem to make the best coolers lately.

I have to say, also, that the GTX 770 is a better card and a better value than the 7970 right now.
 
crashtech said:
The 770 trades blows with the 7970GE, but both seem beyond what the OP is considering. Will the yet unreleased GTX 760 best the 7950 in performance and/or price? Maybe it would be worthwhile to wait and see.
Click to expand...

After realizing that the 7970 wasn't really that much of a boost to the 7950 for the price the 7950 makes more sense. Not to mention new GPU architectures why buy into a 7970 that's been out for how long now? Doesn't make too much sense when next gen gaming will change so much.

I thought about the 760 but I feel like I could get some cash back litecoin mining. Once I factor that in, it really makes little sense to wait. I might as well start mining, try and get $50+ dollars back, and when I sell my 7950, even if I only get 100 dollars, I'll have pretty much (ignoring time value of money) only paid 120 or so. That's not bad for a 7950.

Unless anyone has any other ideas, any other cards that are good price for the 7950 that I should consider or is the one I just posted the best value?
 
Dude, this is the longest running what should i buy thread ever. Just get that Sapphire you linked to already. It's fits your needs. I can vouch for it. I have one. Rock solid, quiet, cool, etc. You can't do much better in price right now, without buying an XFX and sacrificing way too much quality as a result. Others have mentioned worthy alternatives like HIS, ASUS, Gigabyte, etc., but in the end you are talking about very small differences in quality and price. Relax and enjoy life instead of trying to squeeze another $5 out of this decision.
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top