ZOTAC GeForce GTX 650 Ti Boost 2048 MB

[W1zzard]

To read this review go to: http://www.techpowerup.com/reviews/Z..._650_Ti_Boost/

[DarkOCean]

typo on first page it says $150, isnt $170?

[crazyeyesreaper]

So same price to more expensive, and it performed worse with a very tiny overclock. Thats kinda sad lol.

[W1zzard]

Originally Posted by DarkOCean

typo on first page it says $150, isnt $170?

yes, i mistyped this 4 times now .. wishful thinking

Originally Posted by crazyeyesreaper

So same price to more expensive, and it performed worse with a very tiny overclock. Thats kinda sad lol.

but much better noise

[crazyeyesreaper]

True enough, Noise is a big issue for me so in that regard it truly is better but they could have done far more with this card then they did which still makes this rather sad. Oh well not a bad card by any means in terms of noise temps just they could have clocked it a bit higher considering the temps and noise levels are great.

[W1zzard]

had they clocked it higher they'd probably ask more $$ for it

[crazyeyesreaper]

could have clocked it another 7 mhz and made it a "GHz Edition" lol before boost is taken into consideration Still yea higher prices in this segment are a bad thing as you get closer to much better performing parts for a few $$ more So as usual you are right.

[Delta6326]

Great review! This card is tinny and powerful! Sad it's stronger than my 2x 4850 I really need to upgrade...

Do you know why or how it is slightly slower? did they switch around stuff on PCB?

EDIT: re-read the Conclusion.

[Dj-ElectriC]

"Despite OC, card is slightly slower than reference design"

It's funny. it's funny becuase the first thing that came through my mind was "did they even test it to see the performance gains?"

[Ikaruga]

Thanks for the review W1zzard. I'm really curious why the actual - real scenario- performace gains are ended up to be the same despite the different clock values on the three cards, puzzling indeed.

Originally Posted by W1zzard

What I find a bit surprising is that ZOTAC's card performs a little worse than the reference design despite having a 13 MHz overclock.

Just guessing but it's probably a bios issue most likely or perhaps the driver expecting reference boost values atm and that's needs to be tweaked as well.

Just a small suggestion: perhaps it would be better to only include OC-ed values for all the cards in OC-charts like this one:

[crazyeyesreaper]

There is one problem with that Ikaruga, every driver revision would mean that w1zz would not only have to REBENCH ALL THE CARDS, he would also have to Rebench every cards overclock,

no one a chart like that it means W1zz has to run the BF3 benchmark 28 times vs 14 times + 1 oc run, Using older data from previous overclocks at launch would also be well shady lol,

[Ikaruga]

Originally Posted by crazyeyesreaper

There is one problem with that Ikaruga, every driver revision would mean that w1zz would not only have to REBENCH ALL THE CARDS, he would also have to Rebench every cards overclock,

no one a chart like that it means W1zz has to run the BF3 benchmark 28 times vs 14 times + 1 oc run, Using older data from previous overclocks at launch would also be well shady lol,

You right and I know that of course. Well, perhaps I was just too maximalist, sorry.
Maybe doing it against some of the close competitor cards would still come handy, like against OC-ed 7790, 7850, 650ti, 660 in this case. That would only mean 4 more benches for each tests. I understand that I just suggested more work for W1zzard ofc (insta-ban incoming), so it's just a suggestion to make things even better perhaps, after all, this debate/subject pops up in every new card thread, it would help the discussion imho.

[Delta6326]

For me it just seems easier to just go to the other card look at it's OC in a new tab.

Also if you are going to be adding OC results then you have to decide what cards. if say you do a 7790, do you do reference card or the Sapphire or the ASUS or all, it would get very complicated fast.

[douglatins]

i bet the perf issue is well withing the margin of error of tests

[Ikaruga]

Originally Posted by Delta6326

For me it just seems easier to just go to the other card look at it's OC in a new tab.

Also if you are going to be adding OC results then you have to decide what cards. if say you do a 7790, do you do reference card or the Sapphire or the ASUS or all, it would get very complicated fast.

http://tpucdn.com/reviews/Zotac/GeFo...es/perf_oc.gif http://tpucdn.com/reviews/Sapphire/H...es/perf_oc.gif

Yes but those OC values happened with a different driver, so we could end up with misleading data like that, and that's would be very unTPUish

[Sainesk]

How come there's not been Starcraft 2 tests in some recent GPU reviews? I mean the stress it can put on the GPU competes with, if not exceeds that of Diablo 3 (which is arguably a dying game); and a SC2 late game 4v4 match can really push even the best CPUs (is this the reason why? that the CPU after a strong enough GPU would hold back anything stronger?). I just thought it would make a better test than some of the games especially if the test was changed from the usual 1v1 replay.

Appreciate and love your reviews , just curious on knowing more on how and why games to be bench-marked are picked...

[W1zzard]

Originally Posted by Sainesk

How come there's not been Starcraft 2 tests in some recent GPU reviews? I mean the stress it can put on the GPU competes with, if not exceeds that of Diablo 3 (which is arguably a dying game); and a SC2 late game 4v4 match can really push even the best CPUs (is this the reason why? that the CPU after a strong enough GPU would hold back anything stronger?). I just thought it would make a better test than some of the games especially if the test was changed from the usual 1v1 replay.

Appreciate and love your reviews , just curious on knowing more on how and why games to be bench-marked are picked...

SC2 recently had a big update which invalidated all benchmark data. Retesting all cards was not possible in the short timeframe since Heart of the Swarm was released so I had to kick it out.

It will be back, also new will be Bioshock Infinite, Crysis 3 and Tomb Raider, and all cards will use the latest drivers. Starting rebench on NVIDIA next week and AMD as soon as their WHQL is released. The rebenching process will take about two weeks of non-stop testing.

[jagd]

* sends pizza and free beer * , good luck

Originally Posted by W1zzard

Starting rebench on NVIDIA next week and AMD as soon as their WHQL is released. The rebenching process will take about two weeks of non-stop testing.

[W1zzard]

Originally Posted by jagd

* sends pizza and free beer * , good luck

it would be so awesome if people could send me pizza and beer... i would become as fat as gabe newell though

[Delta6326]

Originally Posted by W1zzard

it would be so awesome if people could send me pizza and beer... i would become as fat as gabe newell though

If you lived here in USA and was close to me I would drive the pizza to your home

Awesome news about the re-test! But then again it sounds very boring...

[Ikaruga]

Originally Posted by Delta6326

If you lived here in USA and was close to me I would drive the pizza to your home

Awesome news about the re-test! But then again it sounds very boring...

I spent about 10 mins looking for a Pizza delivery with online ordering options in Stuttgart, where we could order and donate pizzas to W1zzard, when I realized that we don't know the address

[pjl321]

Why is there no SLI benchmarks for this card yet, like anywhere?

[jagd]

LOL'ed , i also looked a place to order pizza and found this http://www.deliver24.de/ but no english . I would like to know if pizza placesdeliver beer in germany ?

Originally Posted by Ikaruga

I spent about 10 mins looking for a Pizza delivery with online ordering options in Stuttgart, where we could order and donate pizzas to W1zzard, when I realized that we don't know the address