Maxwell Gen 1

The specs on the 750 and 750Ti are now official, and somewhat hard to believe.

GTX 750Ti:
640 cores
60W TDP
1020MHz base/1085MHz Boost

GTX 750:
512 cores
55W TDP
1020MHz base/1085MHz Boost

These both use the same 28nm GM107 chip.

Performance? Heh. During Nvidia's conference, they showed a slide comparing frames per second while playing Call of Duty: Ghosts.
The HD 4600 IGP scored 10 FPS
A GT 640 (65W version) scored 22 FPS
And the 750Ti scored 51 FPS

This worries me. The performance increase is simply too good in games. When Nvidia went from the 500 to the 600 series, gaming performance increased alot there too, but the 600 series was worse at crunching. Supposedly these new Maxwell chips in the 750 and 750Ti offer 35% more performance per core while using half the power. That's just too much. Something's wrong somewhere.

When you say, "something's wrong here" do you mean they are trading off Some kind of raw crunching horse power for a certain optimization in games? I'm not sure what that could be, unless it's something like sacrificing details in a scene for higher fps?

Trading off some kind of raw crunching power, yes. Sacrificing game details wouldn't happen though. It's possible Nvidia crippled their FP64 even more, and possibly their INT too. Maybe there aren't even any FP64 units in these cores at all and all INT and FP64 math is converted to FP32 then reconverted to INT or FP64 for output (essentially an emulator). In games you wouldn't notice this at all since they mainly focus only on FP32. In crunching though, it would be horrid.

When someone makes the correct slot configuration of the 750, i'll get one and test it out to see. I suppose there's some slim possibility it really is faster in all areas, i just find it hard to believe at the moment. There's been no die-shrink here. It's the same process used since the 600 series. Maxwell Gen 2 will have the die-shrink to 20nm and i'd be much more likely to believe these numbers then, but without the die-shrink claiming to have 35% more performance per core at half the TDP is a huge leap for a single generation. After the 20nm Maxwells start showing up, how much of an additional performance increase over Gen 1 are we looking at? lol it's just too good to be true.

Here's a comparison - GK107 (Kepler) vs. GM107 (Maxwell)

The GM107 die is 26% larger, has 256 more cores, has 570 million more transistors, 10% more bandwidth, is rated at 494 *more* GFLOPS FP32 (a whopping 61% increase), but uses 6.25% less power to do all this. Oh and 400MHz faster VRAM too (an 8% increase).

So Nvidia is telling me i can crunch 61% faster using slightly less electricity in just 1 generation? I call BS :p

BTW these GPUs are launching in 3 days.

Quote:

---

Originally Posted by John P. Myers

...
So Nvidia is telling me i can crunch 61% faster using slightly less electricity in just 1 generation? I call BS :p

BTW these GPUs are launching in 3 days.

---

Yeah, maybe it is too good to be true, but man that would be nice. :)

I always thought the amount of core's was the most important for crunching, they don't appear to have that many cores for BOINC Crunching ...

GTX 750Ti:
640 cores
60W TDP
1020MHz base/1085MHz Boost

GTX 750:
512 cores
55W TDP
1020MHz base/1085MHz Boost

I think a better energy efficiency card as in a normal energy system is the way to go. Maybe it is not a BS but true. Nvidia is going in the right direction when energy consumption is concerned. For the last 3 years I've been an energy analyst and that's the way people in the area do, they try for the same about of electricity to produce more. Take my hat off to Nvidia if it comes true.

Quote:

---

Originally Posted by STE\/E

I always thought the amount of core's was the most important for crunching, they don't appear to have that many cores for BOINC Crunching ...

---

No they don't have many cores as this is a lower-midrange GPU. There are a few reasons i'm really interested in this though. It's the first chip based on the Maxwell architecture and since it's not a high-end GPU, the performance of the 750Ti will actually dictate how the rest of the Maxwell GPUs (the GTX 800 series) perform. Efficiency always increases as you move to the larger more expensive GPUs and with the 750Ti's GFLOPS/W rating starting out higher than even a GTX 780Ti, it follows that the rest of the 800 series would completely obliterate anything else ever made.

Another thing about the GPUs is you could buy 8 of the GTX 750Ti's for about the price of a Titan but compute 2.32 times faster. Another reason i like looking at these lower-end cards is the lower power draw allows manufacturers to make them only 1 slot wide, and i like finding GPUs i can stick in empty slots just as fillers. Currently this is the fastest, most power efficient GPU that is only 1 slot wide: http://www.newegg.com/Product/Produc...82E16814125468 I was first to review it :D If the 750 and 750Ti are as good as Nvidia says they are, they will become my new #1 choice.

So, Low Profile, which means it will fit in my servers. Looks like a decent upgrade to the 430s that are in there now.

Quote:

---

Originally Posted by Al McAdams

So, Low Profile, which means it will fit in my servers. Looks like a decent upgrade to the 430s that are in there now.

---

Yes, much better, but don't get one just yet. With the GTX 750 rated at only 55W, it should also be possible to make one low profile and *should* outperform the 640 rev2 by better than 25%.

Well, you just stopped me from buying 2. Please keep us update...as if you wouldn't. :)

Quote:

---

Originally Posted by John P. Myers

No they don't have many cores as this is a lower-midrange GPU. There are a few reasons i'm really interested in this though. It's the first chip based on the Maxwell architecture and since it's not a high-end GPU, the performance of the 750Ti will actually dictate how the rest of the Maxwell GPUs (the GTX 800 series) perform. Efficiency always increases as you move to the larger more expensive GPUs and with the 750Ti's GFLOPS/W rating starting out higher than even a GTX 780Ti, it follows that the rest of the 800 series would completely obliterate anything else ever made.

Another thing about the GPUs is you could buy 8 of the GTX 750Ti's for about the price of a Titan but compute 2.32 times faster. Another reason i like looking at these lower-end cards is the lower power draw allows manufacturers to make them only 1 slot wide, and i like finding GPUs i can stick in empty slots just as fillers. Currently this is the fastest, most power efficient GPU that is only 1 slot wide: http://www.newegg.com/Product/Produc...82E16814125468 I was first to review it :D If the 750 and 750Ti are as good as Nvidia says they are, they will become my new #1 choice.

---

But, can you water cool it...? =))

Can we cool it with beer?b-(

Quote:

---

However, double-precision math is further pared back to 1/32 the rate of FP32; that was 1/24 in the mainstream Kepler-based GPUs.

---

$%^$%*! Nvidia

Not very good for crunching. In Europe the 750ti costs the same as the "old" 660GTX,
but the second one has 150% power related to the new maxwell.
For energy purposes the new one is better but both have low power consumption.

plus a question : if I have two ATIs running how do I add a Nvidia in third PCIe port for different projects like Einstein or GPUGrid,
tried this often before but it never worked as I had expected. Used a cc_config for these trials.

If you're using windows vista you can't. Must be xp, 7 or 8. In linux I'm not sure. That's someone else's specialty.

Edit: fp64 has always been horrible from nvidia. People use them for the fp32.

Sent from my Galaxy S4 using Tapatalk Pro

All my boxes run with Windows 7 or 8.
I had the problem that the strong GPUs ran the slow projects ,as Einstein is, after a few hours or days,
The small one made less problems and I was able to fix it to one project.
I could use a slow Ati without any problems but this makes no sense to me.
The third slot should only be used for this half Gpu projects like seti,einstein or wcg.

Quote:

---

Originally Posted by conf

plus a question : if I have two ATIs running how do I add a Nvidia in third PCIe port for different projects like Einstein or GPUGrid,
tried this often before but it never worked as I had expected. Used a cc_config for these trials.

---

No cc_config needed. Just install the Nvidia GPU, install drivers, then select the project you want to run and tell it to only use the Nvidia GPU, then you won't get AMD/ATI WUs from that project. The projects you want to run on AMD/ATI, don't allow them to use Nvidia.



After 3+ hours of digging and digging for real numbers and more info, i found 2 things important to note :)

First is that manufacturers (most likely Zotac and Gigabyte) have stated they will be making true single-slot versions of not only the GTX 750, but of the 750Ti as well. Game on! You can assume from this there will also be a LP version of at least the GTX 750 for Al :p

Second is it *does* appear compute has increased a good bit. I saw FAH Bench results that we proming (Folding at Home's benchmark). I also saw benchmarks in regards to alt-coin mining. Not what we're doing, but math is math. The results were very promising in both cases. When the tru single-slot versions come out, i will buy one that day and get some testing done.

Also thirdly, Nvidia seems to have underclocked these cards big time. People are reporting stable results at 1300MHz with a tiny 31mV bump. EVGA has a version of the 750 with a 1294MHz Boost out of the box (stock is 1085MHz). That's pretty significant.

For crunching, an ATI/AMD GPU MUST occupy the first PCIe slot closest to the CPU. Then the 2nd & 3rd slots can be home to CUDA cards.

So, where are we on this? Any of these new cards single slot?

Quote:

---

Originally Posted by zombie67

So, where are we on this? Any of these new cards single slot?

---

Not yet :/ Believe me, i'm checking daily. I did have a chat with EVGA a couple days ago and asked about it. I was told they won't be making a single slot, which isn't surprising because they never do anyway on the lower cards. Pretty much what i expected. I did find out that the Titan Black will have a liquid version coming soon with the waterblock made by EK.

When the single-slots happen, 99% chance it'll be by Zotac. Maybe 70% chance from Gigabyte.

Edit: The reason it's taking so long is Nvidia was nice enough to redesign the screw hole spacings so any available cooler no longer fits. They're now 44 x 44mm when previously they've been 53.2 x 53.2mm or 58.4 x 58.4mm. So now all the GPU manufacturers have to design new heatsinks from scratch. Nvidia screws us over once again :/

What about AMD? How good is their best single slot GPU?

Quote:

---

Originally Posted by zombie67

What about AMD? How good is their best single slot GPU?

---

The HD 7750 is AMd's best which is about equal to a GT 640 Rev2.

Sapphire Radeon R7 250 ??
http://geizhals.at/sapphire-radeon-r...-a1070502.html
Seems to be equal but is a new card from 2014.

Quote:

---

Originally Posted by zombie67

What about AMD? How good is their best single slot GPU?

---

Quote:

---

Originally Posted by John P. Myers

The HD 7750 is AMd's best which is about equal to a GT 640 Rev2.

---

Quote:

---

Originally Posted by conf

Sapphire Radeon R7 250 ??
http://geizhals.at/sapphire-radeon-r...-a1070502.html
Seems to be equal but is a new card from 2014.

---

Yes, best single slot from the Radeon series will be 7750 or R7 250 (or possibly 6850), but the FirePro series has some way more powerful single-slot cards. Best single-precision is FirePro W7000 with 2432 GFLOPS, best SP/DP card is V7900 with 1860/464 GFLOPS.

XFX FX-775A-ZNP4 Radeon HD 7750 Core Edition

Amazon: $100.74 w/Free Shipping

The FirePros are too expensive anyway. Maybe one can find a used one on Ebay.

Edit. one v7900 found at Ebay for 399,- Euro, not so bad as I thought it would be.

So, are any of these going to be available in, 1) low profile and 2) no need for and additional power connector? I'm looking for the best card to put in my servers to replace my aging 430s.

Quote:

---

Originally Posted by Al McAdams

So, are any of these going to be available in, 1) low profile and 2) no need for and additional power connector? I'm looking for the best card to put in my servers to replace my aging 430s.

---

The XFX unit I linked to doesn't require any PCIe power cables. Found a nice overview vid of that card (Here)

Edit: It looks like there is currently a premium attached to the low-profile versions of the 7750. VisionTek's is $30~$40 more then the XFX single slot.

I assume the low profile kit he mentioned is a low profile kit for that card? If so, that would work. I'll look around and see if I can find anything to give me an idea on the performance increase I might expect over the 430.

Quote:

---

Originally Posted by Al McAdams

I assume the low profile kit he mentioned is a low profile kit for that card? If so, that would work. I'll look around and see if I can find anything to give me an idea on the performance increase I might expect over the 430.

---

I run a low-profile Sapphire HD 7750 in a Dell box - it was great on the old Poem OpenCL app; now have it on Collatz where it puts out about 200k/day. Output on DiRT is comparable.

Quote:

---

Originally Posted by Al McAdams

I assume the low profile kit he mentioned is a low profile kit for that card? If so, that would work. I'll look around and see if I can find anything to give me an idea on the performance increase I might expect over the 430.

---

Me thinks that was just a sales brochure. That XFX card is too tall for low-profile

Besides the one slot design there is a new powerful but cheap Ati card for the third PciE slot:
Sapphire Radeon R7 260X OC, 1GB GDDR5 (115W , 2061GFLOPS (Single), 129GFLOPS (Double)

It costs only 135 $ in Europe but takes two slots.

Quote:

---

Originally Posted by Al McAdams

So, are any of these going to be available in, 1) low profile and 2) no need for and additional power connector? I'm looking for the best card to put in my servers to replace my aging 430s.

---

If you're asking about the GTX 750/750Ti, they are both available without additional power connectors. Looks like it will be a bit yet before any true single slots and low profiles are made though, since Nvidia changed the heatsink screw hole spacings causing heatsinks to be redesigned from scratch.

Galaxy made a low profile GTX 750 Ti :D Only problem is, i can't find it anywhere. Single bracket but heatsink/fan is 2 slots wide. Still it may work in some slim/server cases. http://www.galaxytech.com/__EN_GB__/...se&isPow=False

Edit: nvm found galaxy's online sales. Here it is: http://store.galaxytechus.com/GALAXY...ffer_p_90.html

And the GTX 750 also low profile: http://store.galaxytechus.com/GALAXY...ffer_p_89.html

Cool! That might actually work for me, LP and no 6-pin necessary. I'll need to do some measuring first.

Interesting thread here: http://einstein.phys.uwm.edu/forum_thread.php?id=10613

Very high load with a single task on einstein. Also, they are coming to the conclusion that the Ti is not worth the extra cost or electricity.

Quote:

---

Originally Posted by Al McAdams

Cool! That might actually work for me, LP and no 6-pin necessary. I'll need to do some measuring first.

---

No good, the cooler is too tall. Needs to be a true single slot, low profile and no additional power connector and an upgrade to a GT430. I could possibly mod the server case and use pcie extenders, but that is not my first choice.

Quote:

---

Originally Posted by zombie67

Interesting thread here: http://einstein.phys.uwm.edu/forum_thread.php?id=10613

Very high load with a single task on einstein. Also, they are coming to the conclusion that the Ti is not worth the extra cost or electricity.

---

It is an interesting thread and for Einstein it appears that with current drivers the additional cost of the Ti version isn't worth the added performance. However, based on a small sample, the 750 Ti outperforms the 750 by roughly 15% on GPUGrid, translating to and additional 25-30,000 credits per day. To me, particularly considering the additional 1Gb GDDR5 on the Ti version, that's worth the additional $40 outlay.

Yeah i don't trust Einstein benchmarks at all. For me, my GTX 580 is 3x faster than my Titan....but only on Einstein.