Test of the Real Nvidia Gtx Titan, single-sli-tri Sli, 4 WAY SLI
-
Ask everyone for forgiveness. But I'm having a lot of trouble now installing the OS and Raid on the little computer. And combined with my mood, it's making it feel like a world to be able to put up some decent results.
I'll try to contribute what I can throughout the night.
My apologies, because I know that this mini-review has caused a lot of anticipation among other things because I myself have aroused it. But when things come bad, it doesn't come out right…
Best regards.
Elp3, If you don't feel like it forget about it, family comes first, when you're feeling down and without motivation things don't turn out well, surely your family needs you more than we do.
Cheer up buddy and best regards.
-
Ask everyone for forgiveness. But I'm having a lot of trouble now installing the OS and Raid on the old computer. And combined with my mood, it's making it feel like a world to be able to put up some decent results.
I'll try to contribute what I can throughout the night.
My apologies, because I know that this mini-review has caused a lot of anticipation among other things because I myself have aroused it. But when things come bad, it doesn't come out right...
Best regards.
Compi, tomorrow will be another day, better rest and have a good time with family that's first. Tomorrow there will be more time to do more things and calmly
A hug!
-
Updated with 3D Mark fire strike and OC attempt
-
ACTUALIZO[/COLOR
I've changed my test rig, now I'm on a Rampage IV with a 3930K at 5GHZ
3D Mark Fire Strike, stock cooler:
Here it seems to be going more or less well, even now, when I put one with OC, finally it shows me the GPU-Z that there is a change in the voltage to 1,2V.. the problem? you see for yourselves… the maximum is close to 1200MHZ.. indeed, I have already managed to overclock things to 1200MHZ, but not even joking does it maintain those frequencies despite being a really cool card, here we have it with the fan at maximum, voltage and OC so it doesn't even go over 55º.. according to Nvidia, the card shouldn't lower clocks, however it lowers them in a brutal way, being almost more time at 1070-1090MHZ than close to 1200MHZ.. explanation? well I don't know.. neither TPD nor temp have been exceeded at any time, and yet, the changes have been super-accused...
For now they end like that one who says to arrive, so I haven't had time to do anything. A little patience, and I'll be hanging things up throughout the night..
Salu2.
Have you tried raising the target temperature of the card? It doesn't appear in Afterburner, but apparently this parameter is the one that really controls the behavior of the boost, much more than the TDP itself (which apparently you haven't even exceeded 100%).
The higher the temperature it has at that value, the boost "understands" that it can play more with the high frequency issues with its linked voltages, until it approaches that temperature which is where it would start to make cuts.
Don't cool it too much either, if you've set the fans to manual, leave it in some test with the stock fans so that it has total control over these, to see if the boost is applied as it seems to go in theory. Don't set the target temperature lower, I'm telling you just in case, because that really can be pernicious for performance.
-
-
Have you tried raising the target temperature of the graph? It doesn't appear in Afterburner, but apparently this parameter is the one that really controls the boost behavior, much more than the TDP itself (which you apparently haven't exceeded 100%).
The higher the temperature at that value, the boost "understands" that it can play more with the high frequency issues with its linked voltages, until it approaches that temperature where it would start to make cuts.
Also don't cool it too much, if you've set the fans to manual, leave it in some test with the stock fans so that it has full control over these, to see if the boost is applied as it seems to go in theory. Don't set the target temperature lower, I'm telling you just in case, because that can be really detrimental to performance.
You've hit the nail on the head..you have to let it do TODO.No I like it, but it's how it is..don't touch anything, not even the fan, if you force it, it won't OC the same way..I've put everything on auto with the Guru II which is the one from Gigabyte, and things have gone better..it's in the first post.
Regards.
P.D.
Thanks to Jotole and Praimus…we do what we can, really..a hug.
-
You've hit the nail on the head..we have to let it do TODO.No I don't like it, but that's how ocea..don't touch anything, not even the fan, if you force it, it won't ocea the same way..I've put everything on auto with the Guru II which is the one from Gigabyte, and things have gone better..it's in the first post.
Best regards.
P.D.
Thanks to Jotole and Praimus…we do what we can, really..a hug.
The EVGA precision 4 has already come out, maybe that will work better, Nvidia seemed to have simplified the boost issue for OC and in the end it seems to have messed it up more, although maybe neither the drivers nor the software for OC are adapted for those cards, the precision that has just come out will be for the Titans I suppose.
EVGA | Software | EVGA Precision
regards
-
You've hit the nail on the head..we have to let it do TODO.No I don't like it, but that's how it is with ocea..don't touch anything, not even the fan, if you force it, it won't ocea the same way..I've set everything to auto with the Guru II which is the one from Gigabyte, and things have gone better..it's in the first post.
Regards.
P.D.
Thanks to Jotole and Praimus…we do what we can, really..a hug.
Oks, what you could tweak is the issue of raising TDP and temperature limit values, those are just "instructions" of how the boost should work to be more relaxed.
I'm not sure if it's advisable to raise the voltage, because what it really does is raise 2-3 steps of 13 MHz extra with higher voltages, that is, even if the Oc increases this way, the fact is that the entire scale of previous steps remains the same, with the same voltage, and therefore it shouldn't be useful to achieve higher OCs beyond the addition of those 2-3 extra steps of frequency (30-40 MHz),
The thing with the fan is what I suspected, with everything managed by the Boost itself in temperatures, possibly messing with it there will affect its behavior negatively, it's possible that better control softs will come out soon that take into account the peculiarities of boost 2.0, but for now they seem "rustic" to be applied in Titan without those oddities seen.
The temperature shouldn't matter much with this type of control, so don't be afraid to touch it to "heat up" even though it may seem counterproductive and even "illogical" to look for maximum temperatures. The boost surely operates "more comfortably" seeing itself with margin, and as seen in reviews, basically those maximum temperatures are not reached just like that, so they can be seen as a way to release the barriers self-imposed by the boost.
Temp+TDP at the top and tweak only frequencies to add a generous offset, and rely on the Boost for the extra it manages.
P.D.: A possible way to "bypass" the operation of Boost 2.0 would be the activation in the control panel of the calculation in double-precision floating point (ratio 1/3 instead of the standard 1/24), this option seems to disable the boost technology and return the graphics to a "conservative" mode of operation.
But indirectly it can be good for playing with fully manual OCs, by taking the boost out of the equation if you find it too intrusive (it's not eliminated entirely, the "protection" of TDP values and temps is always activated but what the graphics doesn't do is increase by itself, so surely the frequency shift you have to apply in that mode is much greater for the same thing).
So in this matter it's also advisable to have Temp and TDP at the top, just to prevent the boost from working even at the low end, that is, the frequency dance should be less this way. And maybe you can play with the voltage a bit in that mode, but leave that for later because possibly the gain is little, and it's better to understand the "no voltage" behavior of the graphics first.
Regards.
-
Wow bugs. Thank you very much for sharing it with us.
Best regards
-
Oh my oh my oh my….what a big belly!!!!
-
Man pazuzu what a joy to see you around here, come in and make yourself comfortable.
As for the Titan, of course it's a pleasure to read people like wendigo and majo and what you learn from them.
Friend Elpe I'm really sorry for what you must be going through and I send you a hug, the tests for when you can and have the mood, we'll be here waiting for them.Best regards.
-
Greetings troops. So many familiar faces!
Very green drivers it seems, no?
-
I see that you already have them!! What a blast, I'm eager to see the 3 chicks working, let's see how much they can get out of a tri of 7970 hehe
-
The graphics card is a beast, but there are clearly problems with the boost 2.0 independent of consumption, temps or ventilador.Me it seems incredible that Nvidia can put something like this on the market and on top of that for the price it is…
I tell the story in the main post.
Regards.
-
Who could have 3 of those, what envy you give us!
That said, for the price they are worth, I expected more, much more. Personally, I will wait for the GTX780 which I hope will be better in terms of price/performance than the Titan.
Thanks for the ELP3 tests!
-
Let's see what results it gives with the rampija and that micro, I would like to know your opinion with a GPU since it would be my equipment XD maybe an SLI in the future. -
Let's see what results it gives with the rampija and that micro, I would like to know your opinion with a GPU since it would be my equipment XD maybe an SLI in the future.
It works better on Ivy…but that, I already suspected.;)
Regards.
-
If I had to ask, I'd like a comparison of a single Titan vs a SLI of GTX680, for price equivalence basically; D Have you located any similar comparisons out there? Without surround and maximum 2560 x 1440.
-
I think it should outperform itself with drivers, but even so, the performance/price ratio is disproportionate. They have released some modified drivers from the Titan for the 680s, in which they say that they increase the performance quite a lot, if anyone dares, it would be very interesting to see how much difference the 680s leave from the Titan Download 314.09_Titan_win8_winvista_win7_64bit_english_PreModded.rar from Sendspace.com - send big files the easy way -
What a great job. Thanks Elp3,
With work like this, those of us who come behind with the graphics are helped, even if it means you have to break your head ;D. I really expected something like this but I don't know why, from the reviews I've been looking at, I haven't found any that inspired confidence in me and as you say, many put the "theoretical" functioning of the boost but until today, even Nvidia didn't know there were problems, hehe, this is a joke, but it's just that I haven't read anything about it, I think they haven't gotten into the matter like you, which is what I was afraid of.
Greetings!!
