Upcoming GTX 750 Ti tested

[SiliconWars]

@SiliconWars Its an overclocked IV-E. Exactly the wrong system to measure power consumption in.

Ok granted it's unlikely that either of these cards would find itself in an Ivy-E setup - and frankly I despair of most review sites these days for continuing to benchmark completely unrealistic setups - but let's just swap it out for something like an i5 instead.

So chop off what, 40W or so?

144W for the 750 Ti
190W for the R7 265

49.9 fps for the 750 Ti
63.3 fps for the R7 265

750 Ti = 0.35 = 100%
R7 265 = 0.33 = 94%

So the 750 Ti just eeks out a 6% perf/W victory...but you've still spent the same money on a card that is 20% slower. Not really a great buying decision imo, trading higher performance for lower Wattage.

 

[SiliconWars]

What an argument.

When discussing platform perf/W, yes. System perf/W is the TRUE perf/W.

No, it's not. If you have no headroom for 50-100W more then the GTX750TI is the best option.

50-100W? The R7 295 draws between 30 and 45W more on average. That leaves the 750 Ti advantageous to those people with...I dunno 200W PSU's?

Bingo. The plattform alone is using a huge amount of power. And nobody will combine a GTX750TI with a overclocked 6-Core CPU...

Oh so now we should change the reviewing setups? We could always go back and redo the Mantle benchmarks with an X6 too?

 

[sontin]

You dont get it, huh?

Using a whole system makes worse perf/watt cards much better because the increase in % is so much lower.

So going from 180W to 230W is better for the 265 than from 120 to 170W.

When discussing platform perf/W, yes. System perf/W is the TRUE perf/W.

Oh so now we should change the reviewing setups? We could always go back and redo the Mantle benchmarks with an X6 too?

No, but then i prefer a nVidia and Intel system against AMD and AMD. :lol:
Think about it: Who would combine n power hungry card like the 265 with a Intel CPU? Right. Bulldozer and Radeon it's a much better fit.

 

[SiliconWars]

You dont get it, huh?

Using a whole system makes worse perf/watt cards much better because the increase in % is so much lower.

So going from 180W to 230W is better for the 265 than from 120 to 170W.

Yes I do get it. There are two parts to perf/W but for some reason certain quarters only seem to want to concentrate on the W part.

Let's use your numbers (without adding the extra 4W like you have)

Power

750 Ti: 124W
265: 170W

FPS

750 Ti: 49,9
265: 63,3

Overall

750 Ti = 0.4 = 100
R7 265 = 0.37 = 93

So there you go - a "huge" 7% perf/W victory for the 750 Ti, in about the leanest setup imaginable. For that you lose a flat 20% fps.

It's a good card. It's not an R7 265 competitor except for in a low-power HTPC.

 

[Cloudfire777]

I don`t understand why you people use so much energy in trying to make certain people change their minds:

Here is perf/watt. Maxwell is in a league of its own. 51% more efficient than the 2nd efficient chip, R7 270X.

Power consumption with Maxwell have been reduced by 40% compared to R7 260X, and is still beating it by 15%. Compare it with Nvidia`s own GPUs, GTX 750 Ti is using half of GTX 650 Ti Boost power consumption but is only 8% slower...

This is great news for those who use HTPC. They can use the most coolest stock GPU out there while its only sipping 52W from PCI-e 75W port.
It is also great news for future cards considering that this is 1st gen Maxwell`s. When 2nd gen Maxwell comes in with 20nm, which all mid and high end will do, we will see even greater results.

You need a good foundation to create efficient cards: Wether you try to build a cool GPU, or a quiet one, or have enough room to futher increase performance on future cards. Thats the truth although some closed minded people here try to argue otherwise. Just leave them. There are enough tests that have concluded opposite despite the efforts these closed minded people try to spin the facts.

 

[Silverforce11]

Not sure why you guys are still debating this, Maxwell in the 750/ti is very power efficient, clearly its proven in multiple reviews already.

It is a well crafted chip targets a certain performance range within a low power envelope. If the higher end Maxwell remains as efficient, then it will be awesome.

 

[FiendishMind]

To be fair, the the R7 265 is running quite a bit cooler than the 750 Ti in the review because Ryan's using the numbers from the Sapphire Dual-X cooler R7 265 http://www.anandtech.com/show/7754/the-amd-radeon-r7-265-r7-260-review-feat-sapphire-asus/2

This could easily effect a power usage numbers comparison in the R7 265's favor.

A quick peak at Newegg and I'm not seeing any 750 Ti's using the tiny reference cooler shown in the review.

 

[Abwx]

Power consumption with Maxwell have been reduced by 40% compared to R7 260X, and is still beating it by 15%.

Same score in BF4 though....

http://www.hardware.fr/news/13568/nvidia-lance-geforce-gtx-750-ti-750-maxwell.html

I m not impressed by the power draw, actualy AMD has always
used high voltages for easy bining purposes , binning chips that
can use 15% lower voltage at regular frequency would improve
their power comsumption by 32% , well enough to negate most
of Maxwell s supposed thermal advantages....

 

[Madmick]

When discussing platform perf/W, yes. System perf/W is the TRUE perf/W.

50-100W? The R7 295 draws between 30 and 45W more on average. That leaves the 750 Ti advantageous to those people with...I dunno 200W PSU's?

Oh so now we should change the reviewing setups? We could always go back and redo the Mantle benchmarks with an X6 too?

There's so much focus on dedicated builders here that posters such as yourself sometimes forget that many people out there simply don't share your aims with your system, or are coming at gaming from a different angle.

The GTX 750 Ti and GTX 750 are hands down the best video cards ever produced not requiring external power. Why is this significant? Because even non-gamers or non-builders can add substantial gaming power to their prebuilds. In fact, this saved the life of a friend this week. Talk about opportune timing. His old GT 640 crapped out on him. His PSU was an old Intel prebuild, and it doesn't have the oomph to power higher-end cards; most critically, it doesn't include PCIe connector for powering these cards. It only had a single spare Molex connector (not that I would even attempt to adapt two molex to PCIe on a prebuild). He is an absolute WoW junkie, so his demands aren't high, but the vanilla Intel HD on his CPU can't even handle that. I convinced him to wait a week, and on the 18th advised him to buy the EVGA GTX 750 Superclocked for $125 at Amazon. With this card he can now game in Ultra with 4xAA on 1080p and hold an fps average well above 40 even with his older CPU.

Think about that. Let's consider one context. Without hacking you're not going to change the clocks on your Xbox One or PS4 GPU's, and I suspect that's probably not even a good idea if you could. Meanwhile, even these aggressively pre-overclocked GTX 750 series cards aren't coming with an external port, and thus are therefore achieving this binning without exceeding the 75W ceiling. Did you see the TPU and Guru3D benchmarks for the Palit 750 Ti StormX at 1080p?

That card is whomping on the PS4's rough equivalent in the reference Radeon HD 7850 by nearly ~10%! That's just really attractive to guys who already have one of these AIO comps and just want to be able to play games without spending $500+ on a new system from the ground up.

 

[el etro]

and on the 18th advised him to buy the EVGA GTX 750 Superclocked for $125 at Amazon. With this card he can now game in Ultra with 4xAA on 1080p and hold an fps average well above 40 even with his older CPU.

World of Warcraft is better with Nvidia cards, but for other games you can pick either a HD7850 or a overclocked 260x(260x is faster than GTX 750). Newer Nvidia cards are more about performance/W or compute/w.

That card is whomping on the PS4's rough equivalent in the reference Radeon HD 7850 by nearly ~10%! That's just really attractive to guys who already have one of these AIO comps and just want to be able to play games without spending $500+ on a new system from the ground up.

5% faster than 260x at Hardware Canucks:

Tests by anandtech:

Radeon R7 265 holds a particularly large 19% lead over GTX 750 Ti, and in fact wins at every single benchmark. Similarly, Radeon R7 260X averages a 10% lead over GTX 750, and it does so while having 2GB of VRAM to GTX 750’s 1GB.

 

[Enigmoid]

Ok granted it's unlikely that either of these cards would find itself in an Ivy-E setup - and frankly I despair of most review sites these days for continuing to benchmark completely unrealistic setups - but let's just swap it out for something like an i5 instead.

So chop off what, 40W or so?

144W for the 750 Ti
190W for the R7 265

49.9 fps for the 750 Ti
63.3 fps for the R7 265

750 Ti = 0.35 = 100%
R7 265 = 0.33 = 94%

So the 750 Ti just eeks out a 6% perf/W victory...but you've still spent the same money on a card that is 20% slower. Not really a great buying decision imo, trading higher performance for lower Wattage.

Those are still really high power numbers.

 

[Madmick]

World of Warcraft is better with Nvidia cards, but for other games you can pick either a HD7850 or a overclocked 260x(260x is faster than GTX 750). Newer Nvidia cards are more about performance/W or compute/w.

Yes, duh, dude. Did you not read my post? How is his computer going to power a Radeon 7850 without a PCIe 6-pin power cable connector? How will it power a 260X? It won't. These have a TDP of 130W and 115W, respectively. So they're not an option, and therefore pointless in comparing or considering for his setup. Their power efficiency is precisely what sets the GTX 750 series so far apart for these guys without dedicated gaming builds containing high-powered 12v rail PSUs.

Furthermore, I compared the pre-overclocked Palit GTX 750 Ti StormX operating on the motherboard draw to the equivalent Radeon 7850 in the PS4 which comes at a stock clock, and can't be overclocked (unless hacked, I suppose). The point is that someone in his situation has the option to purchase these higher-clocked GTX 750 Ti cards that outperform, in terms of PC processing power, the reference Radeon 7850. In that same TPU chart, yes, indeed, the reference 7850 outperformed the reference GTX 750 Ti. Obviously consoles and PCs are apples-to-oranges due to software optimization, but I was merely offering the context that in terms of raw processing power he's able to purchase a more powerful GPU than the PS4 correlative and run it on just about any comp in existence.

That's an enormous step forward from the Radeon 7750 and the GTX 650-E which were the previous champions of internally powered cards.

 

[jpiniero]

Yeah, it does make me optimistic that nVidia could do a Denver/Maxwell console (Steam Box?) if they wanted and have something that was very competitive.

 

[Rezist]

I don't get the PS4 comparisons, I mean what the APU uses for power I think is in the 100W or so range, there's no board per say to power. I think think a console will come ahead in power consumption, then there's the raw optimization. Nothing's going to offer the performance of the PS4 either by W or by $.

 

[el etro]

Yes, duh, dude. Did you not read my post? How is his computer going to power a Radeon 7850 without a PCIe 6-pin power cable connector? How will it power a 260X? It won't.

For SFF PCs, yes. For others user just have to use the adapter on a 12v power pin.

 

[Haserath]

Yes I do get it. There are two parts to perf/W but for some reason certain quarters only seem to want to concentrate on the W part.

Let's use your numbers (without adding the extra 4W like you have)

Power

750 Ti: 124W
265: 170W

FPS

750 Ti: 49,9
265: 63,3

Overall

750 Ti = 0.4 = 100
R7 265 = 0.37 = 93

So there you go - a "huge" 7% perf/W victory for the 750 Ti, in about the leanest setup imaginable. For that you lose a flat 20% fps.

It's a good card. It's not an R7 265 competitor except for in a low-power HTPC.

Undervolted mini itx build with SSD and gold+ PSU would be leaner. Could probably get it within 80W with many games.

 

[selni]

Same score in BF4 though....

http://www.hardware.fr/news/13568/nvidia-lance-geforce-gtx-750-ti-750-maxwell.html

I m not impressed by the power draw, actualy AMD has always
used high voltages for easy bining purposes , binning chips that
can use 15% lower voltage at regular frequency would improve
their power comsumption by 32% , well enough to negate most
of Maxwell s supposed thermal advantages....

If AMD reduced their power consumption by 30+% without impacting anything else it'd negate most of Maxwell's power advantage? Well yeah of course, but if they could do that easily they would have already. Stricter binning might help somewhat sure but that's still a massive gap to make up and rejecting more chips would obviously be a bad thing cost wise.

 

[SiliconWars]

Those are still really high power numbers.

Undervolted mini itx build with SSD and gold+ PSU would be leaner. Could probably get it within 80W with many games.

The 750 Ti has been measured to draw just over 60W during gaming. Getting the rest of the system to only draw another 20W seems highly unlikely.

The card still needs at least an i3 to push it otherwise you'd be better off with something like a 7750 instead. That's still at least 40W+ during gaming for that alone, never-mind the rest of the system - 120W is a fair number of what could reasonably be expected of a very low but "balanced" mini-ITX build.

But still, even at 80W what is the difference going to be in reality?

750 Ti: 80W
265: 126W

FPS

750 Ti: 49,9
265: 63,3

Overall

750 Ti = 0.62 = 100%
R7 265 = 0.50 = 81%

That's still just swapping 20ish percent perf/W for the same ultimate performance, and jumping through hoops to get there. Is it worth it? At what point is saving on Watts worth sacrificing 20% performance? Nvidia is ruining their chances by overpricing the card and attempting to sell it as a class above it's true level, simple as that.

 

[SiliconWars]

The 750 Ti will have a clear advantage regarding passive cards and heat being pumped into the case of course - now THAT is something that is a worthwhile advantage in a mini-ITX build.

 

[jpiniero]

I guess the question is whether you would really be able to buy the 265 at $150 or not.

 

[Madmick]

For SFF PCs, yes. For others user just have to use the adapter on a 12v power pin.

Indeed, and that's why I went out of my way to explain that said adapter isn't an option for him.

 

[SiliconWars]

I guess the question is whether you would really be able to buy the 265 at $150 or not.

This, basically. I'm not 100% convinced yet that AMD will sell the R7 265 at $150. You can find the 260X for less than $120 though, regardless of what nearly every review stated.

Indeed, and that's why I went out of my way to explain that said adapter isn't an option for him.

To be frank, there are smarter options than that.

http://www.newegg.com/Product/Produc...82E16817139049

$30 and he'd never have to worry about powering most single chip graphics cards ever again.

What an awesome deal that is btw - note that it comes with a 6+2 pin PCI-E - that should tell you how confident Corsair is about their ability to handle most graphics cards. I've sold a ton of these in the past and never had one back yet.

 

[Abwx]

The card still needs at least an i3 to push it otherwise you'd be better off with something like a 7750 instead. That's still at least 40W+ during gaming for that alone, never-mind the rest of the system - 120W is a fair number of what could reasonably be expected of a very low but "balanced" mini-ITX build.

But still, even at 80W what is the difference going to be in reality?

750 Ti: 80W
265: 126W

Thoses comparisons are moot , actualy the most powerfull card
will likely get the CPU to be maxed out and hence it will inflate
the system power draw because of the CPU, not only because
of the GPU...

 

[Leadbox]

Drivers themselves can have a big impact
I think you'll find nvidia drivers tend to be more cpu heavy
This is why you see at the card power #s looking great and entire system numbers not so great

 

[Haserath]

The 750 Ti has been measured to draw just over 60W during gaming. Getting the rest of the system to only draw another 20W seems highly unlikely.

The card still needs at least an i3 to push it otherwise you'd be better off with something like a 7750 instead. That's still at least 40W+ during gaming for that alone, never-mind the rest of the system - 120W is a fair number of what could reasonably be expected of a very low but "balanced" mini-ITX build.

But still, even at 80W what is the difference going to be in reality?

750 Ti: 80W
265: 126W

FPS

750 Ti: 49,9
265: 63,3

Overall

750 Ti = 0.62 = 100%
R7 265 = 0.50 = 81%

That's still just swapping 20ish percent perf/W for the same ultimate performance, and jumping through hoops to get there. Is it worth it? At what point is saving on Watts worth sacrificing 20% performance? Nvidia is ruining their chances by overpricing the card and attempting to sell it as a class above it's true level, simple as that.

Undervolting can get Haswell down very low(the starting Haswell 3Ghz quad). It might be a bit over.

Not to be too picky, but you're also using a 27% fps difference(from where?- ok, Anand's Crysis 3) compared to 20%.

The performance loss would not be worth it to most builders since PSUs usually start well within the power needed. There are just the special use cases that Nvidia has over AMD.