NVIDIA GF100 & Friends speculation

[mao5]

It's no doubt the same as Tom's Hardwares "sum of all fps" chart

seems correct

 

[neliz]

Vantage Extreme, or?


Another leak (from the 3DCenter-Forum):
http://www.hardwareluxx.de/community/14248999-post364.html

This is interesting too

Die Leistungsaufnahme unter Last ist wie von vielen spekuliert sehr hoch, dürfte aber zumindest Gamer deutlich weniger stören als die Leistungsaufnahme im 2D Betrieb - letztere ist höher als die Leerlaufleistungsaufnahme des gesamten Testsystems inkl. Monitor (aber ohne Grafikkarte).

So, power usage is "very high" under load, but that won't bother gamers. 2D power usage is higher than the complete system including monitor (but without the card.) Must be a driver issue because I thought it had some advanced power saving features.

 

[Lonbjerg]

I guess by "single tests" he means "single game benchmarks".

Perhaps, I am just worried by this bit of his sentence:

synthetic score based on several single tests

 

[GZ007]

Yeah, almost all of them except for the Asus Matrix series. Do SLI, and you need some serious case ventilation to deal with the extra ~360W heatload of twin 4890 GPUs for example...

Arctic cooling's cooler also blocks one extra card slot compared to AMD or NV's reference hardware, clearly not an acceptable alternative to many of us.

The majority of the heat is in the heatsink. U only cool down the fins with air, which enables the heat to move this way again to the fins over and over. The heat always moves from the hotter place to colder.(without any help of course)
If u have to cool down 360W it doesnt mean that u have extra 360W in your case. The heat absorbed with the air in your case is much less (and with more air the temperature is less).

 

[Psycho]

The scores (for the rest of the cards) looks pretty spot on for a balanced average view of the cards - so given it's not fake for the gtx4x0 it should be a quite good measurement..
Some sort of source (ie who posted it where etc) would be nice though..

 

[Grall]

The majority of the heat is in the heatsink. U only cool down the fins with air, which enables the heat to move this way again to the fins over and over. The heat always moves from the hotter place to colder.(without any help of course)

I'm not really sure what you're talking about - and I'm not convinced you actually know either, coz you sound a bit as if you have some english language difficulties.

If u have to cool down 360W it doesnt mean that u have extra 360W in your case.

Watts isn't a unit of heat, it's a unit of power. However, due to resistance electrical power always translates into heat (unless your GPU is superconducting, and that's certainly not the case with the 4890! ), translating into X joules of extra energy dissipated into the air that would not be there otherwise, if the GPU fans had exhausted their cooling air outside the case instead.

So if the heatload of the GPU is dumped into the air inside the case, the case and potentially everything inside it, or at least everything in the airflow path to the exhaust fans if any - and this typically includes the CPU - will heat up a significant amount. I'm not enough of a physicist to attempt any such calculations, but it would absolutely be measureable, particulary in a small case.

The heat absorbed with the air in your case is much less

No idea what you're trying to say, but rest assured, the heat wouldn't simply disappear or anything like that. The heat that leaves the heatsink will be the same amount of heat that absorbs into the case air, in an "open" type GPU heatsink.

 

[Kaotik]

Since the pic is gone (at least for me) I took the liberty of hosting it, too

 

[Silus]

Most people who claim it has 480 SP claim to know it from their own sources, not to choose the rumor they like best. And lately, those people (Fudo, Hardware-Infos, VR-Zone, I think Neliz too) seem to be more numerous than the ones saying 512.

Not really. Both of them already spread both rumors, changing it almost every day. The TDP as well. It was 295, then 275, then 250...And neliz, well...he's been saying that the GTX 480 does NOT have 480 ALUs.

For the record, I didn't say that "if it has 480 ALUs then it will most likely have higher clocks", my point was that whether it has 480 or 512 SPs, it's not going to make a very significant difference, but clocks are.

Disagreed. The architecture allows higher gains in geometry computations, by having more ALUs enabled.

 

[neliz]

Not really. Both of them already spread both rumors, changing it almost every day. The TDP as well. It was 295, then 275, then 250...And neliz, well...he's been saying that the GTX 480 does NOT have 480 ALUs.

Clockspeeds, shader counts and power usage vary by the hour depending on who you ask.


Let's take Fuad.
http://www.fudzilla.com/content/view/18170/1/

Here he is claiming 480CC, why he copied the exact 1401Mhz from VR-Zone is beyond me though, He could've just written 700/1400/900 and be done with it.

It's basically regurgitation of recent rumors. but I believe that my previous statement, about Arun being right (on 480CC) was wrong.

 

[Picao84]

Clockspeeds, shader counts and power usage vary by the hour depending on who you ask.

LOL! I think we never had that much of a silly season, in that sense, before

By the way, nobody noted Neliz post about GF104 tape out some time ago? How come

Anymore info about that Neliz?

 

[Silus]

Clockspeeds, shader counts and power usage vary by the hour depending on who you ask.

Precisely my point
There's no predominant rumor over the other, as someone suggested. People just choose to believe whatever they want. I choose to believe that the highest gains come from having all ALUs enabled, which is exactly what the architecture specs suggest.

 

[Peter Griffin]

5970,5870,5870 2GB 5850 VS. 480, 470 final scores:

195% - gtx480
183% - 5870 1GB
161% - gtx470
153% - 5850

But between the gtx285 and the 5850, difference is high.
Therefore, the tested games a little bit like better Ati.

 

[Alexko]

Not really. Both of them already spread both rumors, changing it almost every day. The TDP as well. It was 295, then 275, then 250...And neliz, well...he's been saying that the GTX 480 does NOT have 480 ALUs.



Disagreed. The architecture allows higher gains in geometry computations, by having more ALUs enabled.

Why would additional ALUs bring higher gains than higher clocks in geometry?

 

[Kaotik]

But between the gtx285 and the 5850, difference is high.
Therefore, the tested games a little bit like better Ati.

Or they just used newer games, in metro 2033 for example the HD5 radeons do a lot better than in older games, when compared to competition (including HD4-series as "competition")

 

[Silus]

Let's take Fuad.
http://www.fudzilla.com/content/view/18170/1/

Here he is claiming 480CC, why he copied the exact 1401Mhz from VR-Zone is beyond me though, He could've just written 700/1400/900 and be done with it.

It's basically regurgitation of recent rumors. but I believe that my previous statement, about Arun being right (on 480CC) was wrong.

Ninja edit I see

Anyway, even VR-zone...they seem to just have copied what Scott believed could be the clock speeds (for both the core and ALUs) in their GF100 architecture preview:

http://techreport.com/articles.x/18332/5

They just used rumors and speculation to spread another rumor...

 

[nagus]

70mm (5850) vs 73mm (470) ~ +5%

83mm (5870) vs 88mm (480) ~ +6%

according to the leaked chart above

 

[Peter Griffin]

Or they just used newer games, in metro 2033 for example the HD5 radeons do a lot better than in older games, when compared to competition (including HD4-series as "competition")

Yes, its possible

 

[DegustatoR]

both of them already spread both rumors, changing it almost every day. The tdp as well. It was 295, then 275, then 250...and neliz, well...he's been saying that the gtx 480 does not have 480 alus.

480. 250.

 

[Silus]

480. 250.

 

[rpg.314]

The architecture allows higher gains in geometry computations, by having more ALUs enabled.

Could you explain in detail, how exactly does this theory of yours work?