Query: Does the 5900U shift its 2D-3D clock speed?

kyleb said:
well sure if we want to go off into fantasy world and exclude 2d mode from the definition of "normal operation"; then you are correct. do you want a cookie Xmas?
Would you call that thermal clock throttling?

Yeah, I like cookies :D
 
no but the nv3x does downclock itself during normal operation, unlike the p4. hence our conversation has completed a full circle. :?
 
I seen the heatsink bottom of a 5600U melted.
I forced it to run some 2d benchmark with a core speed of 1GHz. :LOL:
Shop owners are very kind. ;)
 
kyleb said:
no but the nv3x does downclock itself during normal operation, unlike the p4. hence our conversation has completed a full circle. :?
Yes it does downclock in 2D mode. But the reason I compared it to P4 was that it downclocks when *overheating*.
2D clock throttling isn't relevant here.
 
wtf Xmas, this is absolutely absurd! 2d clock throttling damn well is relevant because that is what WaltC was asking about in the first place. then came the proposal that the 2d throttling could be indicative of the chip not being completely stable if kept permanently at 3dspeeds. to this you responded:

Xmas said:
Try to apply the same logic to Pentium 4 overheat protection.

...to which i replided:

kyleb said:
but the p4 never uses that in normal operation, you have to take the heatsink off or something first.


...and then you went off into your own world from there which there is no point in rehashing as that was all down hill. so is there anything clicking yet for you or is someone going to have to draw out a flow chart? o_O
 
kyleb said:
wtf Xmas, this is absolutely absurd! 2d clock throttling damn well is relevant because that is what WaltC was asking about in the first place. then came the proposal that the 2d throttling could be indicative of the chip not being completely stable if kept permanently at 3dspeeds. to this you responded:
Come on, Kyle. The clock throttling was clearly devised as a way to reduce the noise of the FX 5800 Ultra. That it is included as a feature in the entire FX line should be considered a blessing, not a curse.

Anyway, if any of these graphics cards aren't completely stable under prolonged 3D usage, there will be quite a number of complaints from the 3D enthusiast crowd (and lots of returned video cards).
 
kyleb said:
wtf Xmas, this is absolutely absurd! 2d clock throttling damn well is relevant because that is what WaltC was asking about in the first place. then came the proposal that the 2d throttling could be indicative of the chip not being completely stable if kept permanently at 3dspeeds. to this you responded:

Xmas said:
Try to apply the same logic to Pentium 4 overheat protection.

...to which i replided:

kyleb said:
but the p4 never uses that in normal operation, you have to take the heatsink off or something first.

...and then you went off into your own world from there which there is no point in rehashing as that was all down hill. so is there anything clicking yet for you or is someone going to have to draw out a flow chart? o_O
I think you really need to read this thread again.

WaltC showed his scepticism about the clock throttling feature because he thinks it might indicate that the chips don't run completely stable at 400MHz in 3D, right.

I answered that I don't agree with that and think that this is simply (a) a "silence" mode in 2D operation and (b) a protection feature, included because it was easy to do so, not because it is needed.

YeuEmMaiMai then stated that clock throttling is an indication that the design is not suited for continuous operation at 400MHz.

To which I responded that the Pentium 4 also has overheat protection and clocks down when getting too hot. Still it is suited for continuous operation at its nominal clock rate.

Or are you saying that you believe the NV3x clocks down in 2D mode because it would overheat otherwise??? *That*'s completely absurd.


I mentioned the P4 because it's another chip with thermal clock throttling, clock throttling that is triggered by a too high temperature that is, also known as overheat protection. Then you go on and talk about downclocking in 2D mode and that the P4 doesn't clock down when idling (but I wish there would be a desktop CPU that did!), which has nothing to do with why I brought up that example in the first place.
 
ok, so now you say thermal clock throttling is relevant to the discussion and the p4 is "another chip" that uses it; but the p4 doesn't downclock while idling so that makes me wrong when i was the one saying that exact fact is what makes the comparison invalid. wtf :!: :?: :!:

and no i am not saying i believe that the chip will overheat without under clocking 2d mode, only agreeing with the possibility that others presented above. is this some sort of language barer thing? i ask because this whole conversation seems rather convoluted. :?
 
Just so we're all talking about the same thing:
1) The chip down clocks and fan powers off when running 2d stuff.
2) This doesn't indicate that there's a problem running at speed
3) Quiet is a good thing

Did I miss anything? no? Good. Now lets end the discussion.
 
Just a side comment, the P4's thermal threshold is high enough that, running at stock frequency, the threshold will never be reached, as long as the cooler is at least up to Intel's spec... but it's still low enough that the core will never be in any real danger of a massive burnout (though electromigration is still a concern if you're OC'ing right up to the threshold).
 
RussSchultz said:
Just so we're all talking about the same thing:
1) The chip down clocks and fan powers off when running 2d stuff.

Yes, the gpu clock is cut way down in 2D so that the fan may be turned off until 3D processing is resumed.

2) This doesn't indicate that there's a problem running at speed

No, but the thermally activated clock throttle definitely *might.* (As we saw happen with some reviewed nv30 reference designs. Because this is legacy technology carried over from the original nv30 reference design I do not trust that its presence is completely unnecessary at stock operational speeds in all conditions.)

3) Quiet is a good thing

Relatively speaking, of course--relative to what sort of decibels the fan puts out while 3D processing--because you lose the "quiet" when running 3D.

The point to the original fan speed throttling in the 5800U was not that "quiet is a good thing"--if you mean silent--because the original design simply slowed the fan and clocked down 2D speeds to allow that--so as to contrast to the horrible racket the fan made while running at speed doing 3D processing. Turning the fan completely off while dropping the 2D clock to facilitate that merely allowed some OEMs to advertise a lower average decibel ouput than was possible when the fan was still running in 2D, just slower than when 3D processing.

Relative also to what other fans one may have running at the same time, because a user may not hear a difference when the gpu fan shuts down in 2D contrasted with his other system fans.



Did I miss anything? no? Good. Now lets end the discussion.

Right. Let's...;)
 
kyleb said:
ok, so now you say thermal clock throttling is relevant to the discussion and the p4 is "another chip" that uses it; but the p4 doesn't downclock while idling so that makes me wrong when i was the one saying that exact fact is what makes the comparison invalid. wtf :!: :?: :!:
I compared NV3x and P4 specifically with regard to thermal clock throttling. So why is that comparison invalid?
 
Back
Top