GTX 750 Taken Apart, Sips Power from a Single 6-pin Connector

[alwayssts]

Not to sound like a AMD fanboy but I just read about the rumored AMD R9 265 which looks better than this (specs/price weise) but nothing is certain.

Whomever designed Pitcairn deserves a medal...gotta wonder how much of their profit over the last 2 years came from that chip. I bet it is a huge, huge chunk.

Not too tough to estimate a rough guess, though.

Scary math ahead.

(5400*128)/8 = 86.4gbps (bandwidth)
86.4/56.25 = 1.536TF (max usable shader compute)
1.536TF/960*2 = 800mhz
1150 (boost clock)/800*.16666 (special function) = ~.24
1.24*800 = 992mhz
992*960 = 1.9TF
1.9TF/(1024*2) = 930mhz (265 clockspeed is probably '925mhz' from amd, 975 AIB like 270)
1.9*56.25 = 107gbps
107*8/256 = 3348mhz (using 930mhz)
4800/3348 = 1.433
.433*.16 = ~6-7%

Obviously that number changes a little when considering slight efficiency differences (like tex units, which could bring it down a few percent), use cases (compute would be closer to the raw difference), and the core clock could be a little slower to a quite a bit faster than 930mhz, which while when faster increases performance more linearly than bw, would throw the difference bw makes off as obviously it would require more.

Yeah, it'll be faster at stock...but pretty close while being a much larger chip (even if salvage and using cheap 5ghz ram). That said, more logic + ideally low clocks (~900mhz/<5000mhz, where the process and mem controller are aimed) should keep power in check. The fact amd sells cheap and (usually) nvidia overcharges should keep the prices pretty close. It really comes down to two things:

1. How high can the mem clock on GM107, because while special function/bw voodoo of 16-17% will help at a higher clock when bw is scarce, it will be equally disproportionate scaling if bandwidth cannot feed the shaders.

2. How high can the core scale on 265? If limited to (for example) 1050mhz like 270, and the memory is crap elpida and hits (perhaps) 5300mhz, and the gm107 uses 6ghz on it's superior memory controller reaching closer to 7ghz overclocked, in theory around 1200mhz (little less on both counts) they could be roughly equal. How that plays out in power consumption considering die size, ram amount, etc is a good question.

Would that outcome surprise anyone? I bet it would not...two pretty similar-performing parts overall, even if massively different.

 

[ensabrenoir]

just ordered a 650 ti ti go in my mini itx build . .... would've love to give this guy a spin though.......depends on how review goes though......maybe i might still

 

[Casecutter]

All I'm going to say is... they couldn’t find an uglier spot to place that 6-pin! All the way over by the bracket, I can't say I've ever seen that? So now you’ll have this wire loom running across the card right in front. Perhaps with the prevalence bottom mount PSU I suppose it hardly matters.

We might well see "Ti reference" models with no 6-pin, more a way that AIB's can then substantiate/differentiate their more custom OC'd versions. Go with reference and be assured it only ever be a "plug and play" card, while then the FTW, AMP, Super Clocks, etc offer the bump that’s so lucrative for those AIB’s.

 

[TheinsanegamerN]

The fact that needs a 6 pin power connector eliminates most expectations about a super efficient maxwell architecture. There was a leaked preview that was saying about no 6pin power connector in either card, 750 and 750Ti. It seems that that was a false preview, a lie, maybe a paid article to create a little more hype about Maxwell.
I just lost almost all interest about Maxwell.

why did people expect that, exactly? it is built on the same process, and kepler was already efficient. the big enhancements wont come until the 20nm die shrink at the end of the year. expectations havent been eliminated, they havent even been tested yet.

 

[xorbe]

I can't remember the last time I saw so much hissy fit over a simple low end video card!!

 

[Xzibit]

All I'm going to say is... they couldn’t find an uglier spot to place that 6-pin! All the way over by the bracket, I can't say I've ever seen that? So now you’ll have this wire loom running across the card right in front. Perhaps with the prevalence bottom mount PSU I suppose it hardly matters.

We might well see "Ti reference" models with no 6-pin, more a way that AIB's can then substantiate/differentiate their more custom OC'd versions. Go with reference and be assured it only ever be a "plug and play" card, while then the FTW, AMP, Super Clocks, etc offer the bump that’s so lucrative for those AIB’s.

Well since these cards don't look to be multi-card "SLI" capable like there competition they have that room and power savings to spare.

 

[HumanSmoke]

They also said The ASUS 750 is OC model but does not need a 6pin but the Gigabyte does. Just adds to the confusion.
The PCB from OP looks more inline with the Gigabyte and not the ASUS one. None look like the Asus.

There are only two possibilities:
1. The reference specification does not require auxiliary power, and AIB's are at liberty to add at their discretion, or the extremely unlikely,
2. The reference card requires auxiliary power and some AIB's are manufacturing cards with power circuitry LOWER than Nvidia's reference design
Which would you think is more a more likely state of affairs?

I'd also add that, while adding a 6-pin auxiliary power input is for some reason deemed a crime against humanity by some people, isn't it better that the vendor is allowing custom implementations from day one?
All in all, except for the immediate availability it doesn't seem very far removed from, say, the HD 7970/GE (for example) - there are more than a few 2 x 8-pin varieties of proprietary designs, and the reference board even has solder points for 2 x 8-pin input, yet I never saw a whole lot of howling about the need for 375W input power for the card from the same people who see the incorporation of feature set expansion options with this card as a negative. How curious.

 

[jihadjoe]

If GM107 is Maxwell shouldn't they have branded this the 850?
700 series is going to be hella confusing with Kepler parts filling in the top and Maxwell parts at the bottom.

 

[Xzibit]

There are only two possibilities:
1. The reference specification does not require auxiliary power, and AIB's are at liberty to add at their discretion, or the extremely unlikely,
2. The reference card requires auxiliary power and some AIB's are manufacturing cards with power circuitry LOWER than Nvidia's reference design
Which would you think is more a more likely state of affairs?

I'd also add that, while adding a 6-pin auxiliary power input is for some reason deemed a crime against humanity by some people, isn't it better that the vendor is allowing custom implementations from day one?
All in all, except for the immediate availability it doesn't seem very far removed from, say, the HD 7970/GE (for example) - there are more than a few 2 x 8-pin varieties of proprietary designs, and the reference board even has solder points for 2 x 8-pin input, yet I never saw a whole lot of howling about the need for 375W input power for the card from the same people who see the incorporation of feature set expansion options with this card as a negative. How curious.

I think there are a few other possibilities.

ASUS might be re-using one of their 3 650 PCB or its photoshoped since the cooler looks nothing like a Asus design

Either way. It will be like the 650. Some had 6pin connectors some didn't. Nvidia recomemded one even when it was set at 64w TDP

 

[HumanSmoke]

If GM107 is Maxwell shouldn't they have branded this the 850?

Then the forums would be awash with comments along the lines of "What happened to the 750? I bet it was cancelled", "The 850 is barely faster than the 650, how dare Nvidia try to rip me off even though I'd never buy it whatever the model number is", and "Why bump up it up to a whole new series when the architecture is similar to Kepler and/or it's still on 28nm"

700 series is going to be hella confusing with Kepler parts filling in the top and Maxwell parts at the bottom.

That's what happens when the foundry process node cadence isn't in lockstep with GPU architecture launches. You think it's any worse than AMD's R7 / R9 naming farrago? How many average consumers could differentiate new, refreshed, or rebranded GPUs within R7 240, R7 250, R7 250X, R7 260, R7 260X, R7 265, R9 270, R9 270X, R9 280, R9 280X, R9 290, and R9 290X ? What precisely is the difference between Tahiti XT and Tahiti XTL or XT2?, or Pitcairn and Curacao ?
Welcome to the synthesis of marketing and random number generation.

 

[Arjai]

Conjecture? Anyone?

 

[Patriot]

1. Calm yo tits... wait for 3rd party benches and power readings.
2. It is common that if a card is close to a power bracket edge to include the next size up.
Without it Overclocking would suck and low end motherboards would be strained etc ...

 

[john_]

why did people expect that, exactly? it is built on the same process, and kepler was already efficient. the big enhancements wont come until the 20nm die shrink at the end of the year. expectations havent been eliminated, they havent even been tested yet.

If you where saying this one week earlier I would say that in the end you where right. Saying it now it is just pointless. Today I can also make the same question.
Anyway if you want an answer you should see all the (wrong) info about the cards that was all over the internet the last 10 days pointing at a 960 cores card with no 6pin connector. A month ago when info was very limited I was expecting a card that wouldn't be much different than Kepler because of that 28nm process. The last 10 days wrong info was spread and maintained on the internet with multiple articles, sources, gpuz screenshots and pictures of the cards, so the expectations where totally different. In the end it seems we will end up somewhere in the middle.

 

[Ruru]

Who even cares if a card has power connectors?

 

[HumanSmoke]

Who even cares if a card has power connectors?

Very few, but the inference is that if the card has a 6pin connector it is because it needs a 6-pin connector. The next jump in the faulty logic is, if the card has 150W input power (75W from the slot + 75W from the 6-pin) then it must need close to that to operate...and the whole point of these kinds of GPUs is one of efficiency- always a good talking point on a new architecture. For instance, some people make an assumption based upon an assumption...

That new die size is much better suited to have ROI than the GK106 ever was for them! The 750 will be the 75W part, while the 750Ti could be as high as 110W..

...which not only ends up 36% to 83% off the mark...

...but then is used as a basis for screeds of supposition on a supposed lack of efficiency for a GPU whose raison d'être is efficiency. Of course when it is the same people being overly pessimistic regarding one vendor, and wildly optimistic regarding another, it becomes kind of an armchair sport guessing how far off the mark their next pronouncement is, and how long they keep to their original position in the face of mounting contrary evidence.

[source]

 

[john_]

With Maxwell it seems that whatever you say it will be wrong an hour latter.
60W for 750Ti at 28nm is really a huge step forward.

 

[racedaemon]

I find it strange that this card has no SLI connector as the 650Ti BOOST had. Ar they keeping that feature for a 750 Ti Boost? Was that to much future proofing/performance potential for that price point? Should i go buy a 650Ti BOOST? Only one day to wait for the answer to the last question.

 

[xorbe]

I find it strange that this card has no SLI connector as the 650Ti BOOST had.

The defunct 650 Ti Boost was a gimped 660 (which has SLI).

 

[HumanSmoke]

Looks like some folks got a jump on the official launch

1200+ MHz core / 6400 MHz effective memory with no auxiliary power input. Not too shabby.

[Source]

 

[Xzibit]

It's great that it has no 6pin and performance per watt should be good

Whats troubling is the first picture. Look at the price $199.99 OUCH!!!!

With the R7 260X OC going for $119-$129. A $70 more then 50% difference is too high.

 

[xorbe]

I wonder if "sips power" is the new catch phrase that will be present in every single review.

Good eye on that $199 price, yikes!

 

[racedaemon]

Whats troubling is the first picture. Look at the price $199.99 OUCH!!!!

Let's hope those are not US dollars... although the user that posted the pictures appears o be form the US...

 

[HumanSmoke]

With the R7 260X OC going for $119-$129. A $70 more then 50% difference is too high.

Nice try, but no banana.
The card was purchased from Micro Center - not the cheapest outlet for graphics- an R7 260X will cost you $145-155

 

[Xzibit]

Nice try, but no banana.
The card was purchased from Micro Center - not the cheapest outlet for graphics- an R7 260X will cost you $145-155

Launch prices always match or are better at MicroCenter

MicroCenter OH - Cincinnati/Sharonville - MSI GTX750TI OC TF 2GB D5 DVH $179.99 w/$20.00 discount

 

[HumanSmoke]

Launch prices always match or are better at MicroCenter
MicroCenter OH - Cincinnati/Sharonville - MSI GTX750TI OC TF 2GB D5 DVH $179.99 w/$20.00 discount

Well, that comparison looks pretty much OK, and a lot less dire than:

With the R7 260X OC going for $119-$129. A $70 more then 50% difference is too high.

$180 for the 750Ti OC versus $145-155 for the R7 260X from the same outlet. I guess we'll find out in a day or so what the MSRP is of the majority of cards

I wonder if "sips power" is the new catch phrase that will be present in every single review.

Seems apropos considering the owner of the card said that his i3 / 750 Ti system is pulling less than 120 watts when gaming