UPDATETITAN Z Base 705mhzBoost 876mhz

[Hitman928]

Not as hot as Nvidia s VRMs wich according to hardware.fr are substancialy hotter than on AMD s cards with temperatures reaching up to 120°C, on the 290X review they explicitely say that Nvidia s 780ti power supply is heating more as it was not really designed for extreme overclockings.

http://www.hardware.fr/articles/912-4/bruit-temperatures.html

Where are the VRM's reaching 120 deg? Am I missing it in that link?

 

[Abwx]

Where are the VRM's reaching 120 deg? Am I missing it in that link?

It was in another of their review, IIRC the specified temp was 125°C,
i ll look for the exact article.

 

[Cookie Monster]

It was in another of their review, IIRC the specified temp was 125°C,
i ll look for the exact article.

Techpowerup I think it was. They have a video capture of the thermal imaging, something new they implemented in their reviews.

 

[24601]

Techpowerup I think it was. They have a video capture of the thermal imaging, something new they implemented in their reviews.

You're going to have to link me, because I can't find what you speak of here:

http://www.techpowerup.com/reviewdb/Graphics-Cards/NVIDIA/GTX-780-Ti/

 

[Abwx]

Techpowerup I think it was. They have a video capture of the thermal imaging, something new they implemented in their reviews.

I m sure it was HFR, Damien Triolet the GFX reviewer always do extensive measurements including the VRMs, he pointed that despite the apparently high temp the components where actualy well within safe area and indeed VRMs, mosfets actualy, can withstand silicon temperature of 175-200°C depending of the manufacturing process and mosfets geometries.

 

[24601]

No link still.

Talking out of your ass, Gotcha.

 

[Stuka87]

No link still.

Talking out of your ass, Gotcha.

What good would a link do? You just stated every review site but one is lying, and are just payed shills.

 

[thesmokingman]

I m sure it was HFR, Damien Triolet the GFX reviewer always do extensive measurements including the VRMs, he pointed that despite the apparently high temp the components where actualy well within safe area and indeed VRMs, mosfets actualy, can withstand silicon temperature of 175-200°C depending of the manufacturing process and mosfets geometries.

 

[Abwx]

Yes, macro mosfets are indeed more rugged than nano fets, notice that the 175-200°C are for the silicon not the component case wich will be at about 150°C and even higher when thoses die temp are reached.

You can see in the doc below, operating and storage temp up to 175°C :

http://www.infineon.com/dgdl/Infine...90004&fileId=db3a304412b407950112b40b62ba07b4

 

[24601]

Yes, macro mosfets are indeed more rugged than nano fets, notice that the 175-200°C are for the silicon not the component case wich will be at about 150°C and even higher when thoses die temp are reached.

You can see in the doc below, operating and storage temp up to 175°C :

http://www.infineon.com/dgdl/Infine...90004&fileId=db3a304412b407950112b40b62ba07b4

Wish people could be permabanned for lieing out of their asses.

Have fun running your GPU for 1 microsecond at a time.

 

[Abwx]

Wish people could be permabanned for lieing out of their asses.

Have fun running your GPU for 1 microsecond at a time.

I m talking of VRMs max temp not of the GPU one, isnt that clear.??.

So people should be banned because you re unable to read accurately their sayings and then you go as far as branding me a liar because you dont understand what i m talking about.?..

 

[24601]

I m talking of VRMs max temp not of the GPU one, isnt that clear.??.

So people should be banned because you re unable to read accurately their sayings and then you go as far as branding me a liar because you dont understand what i m talking about.?..

The most critical value for the aging of the MOSFET is
the junction temperature. This
value is also given on the first page and may not be exceeded under any circumstances.
With respect to the previous explanations the junction temperature doesn't exceed the
given maximum rating of 175
̊C. This can be explained by the relatively low dissipated
energy of
only 8.75
m
J
(8.75 kW, 1 μs)
in a very short period of time. The reason for the
ability of the MOSFET to absorb this amount of energy without suffering can be found in
the thermal capa
citance of the device itself.

Learn to read or gtfo.

 

[3DVagabond]

The most critical value for the aging of the MOSFET is
the junction temperature. This
value is also given on the first page and may not be exceeded under any circumstances.
With respect to the previous explanations the junction temperature doesn't exceed the
given maximum rating of 175
̊C. This can be explained by the relatively low dissipated
energy of
only 8.75
m
J
(8.75 kW, 1 μs)
in a very short period of time. The reason for the
ability of the MOSFET to absorb this amount of energy without suffering can be found in
the thermal capa
citance of the device itself.

Learn to read or gtfo.

No reason to be uncivil. We are simply sharing information here.

The reason I mentioned what the VRM temps are likely to be (+100°C) is because everyone has lately been so concerned about temps in that range. Reality is cards that draw +250W or so, run those temps or even higher. It's simply a fact. Nothing to get all bent out of shape about.

Hardware.fr has for a long time now studied temps of both the GPU's and the PCB components. I would advise taking some time, use Google translate if needed, and study their results. You will be surprised how many cards have had +100°C VRM's and have handled O/C'ing with extra voltage to very high levels (The GTX-460 for one).

Too many sites simply run 3 minute benches for their performance results. Hardware.fr and [H] are two sites that actually test cards that have been run until their temps stabilize. The difference in performance sometimes is quite large. It's not just reference 290's that throttle. For some reason though, there has been a large effort to expose their throttling tendencies while ignoring it for most other cards. Just like the large effort to report on the VRM temps.

 

[24601]

No reason to be uncivil. We are simply sharing information here.

The reason I mentioned what the VRM temps are likely to be (+100°C) is because everyone has lately been so concerned about temps in that range. Reality is cards that draw +250W or so, run those temps or even higher. It's simply a fact. Nothing to get all bent out of shape about.

Hardware.fr has for a long time now studied temps of both the GPU's and the PCB components. I would advise taking some time, use Google translate if needed, and study their results. You will be surprised how many cards have had +100°C VRM's and have handled O/C'ing with extra voltage to very high levels (The GTX-460 for one).

Too many sites simply run 3 minute benches for their performance results. Hardware.fr and [H] are two sites that actually test cards that have been run until their temps stabilize. The difference in performance sometimes is quite large. It's not just reference 290's that throttle. For some reason though, there has been a large effort to expose their throttling tendencies while ignoring it for most other cards. Just like the large effort to report on the VRM temps.

290 is ~370W TDP @ 947 Mhz, 290x is ~ 420W TDP @ 1000 Mhz.

The GPU is fed through a 5 phase VRM setup, rated at 50 amp @ 100C(core).

50 x 5 = 250w @ 1v, 250 x 1.17 = 292.5w @ 1.17v (Typical droop from stock voltage of 1.25v @ max load)

That means the VRMs that feed the GPU on the card can only take ~300w @ 100C.

This is the reason why 290s and 290x's crash so often when they are at max load @ default clocks when you turn off all throttling if you allow the VRM to get over 80C.

I don't know how often I have to explain the most basic things here, but I've done this at least 20 times so far since before the 290 and 290x even launched.

I did the same explanations with the 7970 and 7950.

I'm sick and tired of people questioning me when I'm obviously right with mathematical, empirical, and anecdotal data all lined up.

I've literally given more evidence than a doctoral dissertation of a research scientist.

And people still go with "nuh oh, magic AMD/ATi no can has do wrong brah" as their evidence in their corner.

 

[Abwx]

The most critical value for the aging of the MOSFET is
the junction temperature. This
value is also given on the first page and may not be exceeded under any circumstances.
With respect to the previous explanations the junction temperature doesn't exceed the
given maximum rating of 175
̊C. This can be explained by the relatively low dissipated
energy of
only 8.75
m
J
(8.75 kW, 1 μs)
in a very short period of time. The reason for the
ability of the MOSFET to absorb this amount of energy without suffering can be found in
the thermal capa
citance of the device itself.

Learn to read or gtfo.

This just prove that you re deliberatly creating polemics because i mentionned silicon temperature wich is obviously junction temperature isnt it but i guess that your point is not accuracy.

The VRMs silicons, junction if you prefer, can withstand at least 175°C permanently temp as written in all manufacturers datasheets, it s 20 years that i m using those components and reading through tec docs and it is obvious that you dont know what you re talking about as thermal response is just another matter to help define SOA in burst thermal stress.

Besides if a manufacturer say 175°C he will forcibly include a safety margin, other than that it seems that all your problems lies with my statement that Nvidia VRMs run much hotter than in AMD cards....

 

[3DVagabond]

290 is ~370W TDP @ 947 Mhz, 290x is ~ 420W TDP @ 1000 Mhz.

The GPU is fed through a 5 phase VRM setup, rated at 50 amp @ 100C(core).

50 x 5 = 250w @ 1v, 250 x 1.17 = 292.5w @ 1.17v (Typical droop from stock voltage of 1.25v @ max load)

That means the VRMs that feed the GPU on the card can only take ~300w @ 100C.

This is the reason why 290s and 290x's crash so often when they are at max load @ default clocks when you turn off all throttling if you allow the VRM to get over 80C.

I don't know how often I have to explain the most basic things here, but I've done this at least 20 times so far since before the 290 and 290x even launched.

I did the same explanations with the 7970 and 7950.

You simply ignore anything that doesn't fit in with what you believe. I'm not even sure where you get your figures from since you post no sources. Look at what else has been said and shown by other posts and documented by respectable sites. You can believe that it's some paid conspiracy theory if you'd like to.

 

[24601]

You simply ignore anything that doesn't fit in with what you believe. I'm not even sure where you get your figures from since you post no sources. Look at what else has been said and shown by other posts and documented by respectable sites. You can believe that it's some paid conspiracy theory if you'd like to.

You deny the very document that your compatriot linked?

That's some convenient compartmentalization you have going on there.

I've disproved your guys' crap for over 1000 posts now and I have been proven correct by time 100% of the time.

You wanna bet on you being correct?

Hell, I even got the actual TDP of the 290 and 290x and the 295x2 down to a few percentage points >>BEFORE THE CARDS WERE EVEN SOFT-LAUNCHED<<

According to your guy's information, I've beaten both the market and first party predictions for earnings trends for NVIDA and AMD/ATi for 3 quarters now.

According to your guy's information, I've absolutely trounced both AMD/ATi and Nvidia's entire competitive analysis departments for over 1 year now.

 

[Abwx]

The GPU is fed through a 5 phase VRM setup, rated at 50 amp @ 100C(core).

50 x 5 = 250w @ 1v, 250 x 1.17 = 292.5w @ 1.17v (Typical droop from stock voltage of 1.25v @ max load)

That means the VRMs that feed the GPU on the card can only take ~300w @ 100C.

This is the reason why 290s and 290x's crash so often when they are at max load @ default clocks when you turn off all throttling if you allow the VRM to get over 80C.

These calculations means absolutely nothing, just taking 1V output voltage instead of 1.2 will reduce the calculated power by 20%, when i see such blatant and willfull miscalculations i dont even bother reading the rest.

 

[24601]

[QUOT=Communism;36313768]
The GPU is fed through a 5 phase VRM setup, rated at 50 amp @ 100C(core).

50 x 5 = 250w @ 1v, 250 x 1.17 = 292.5w @ 1.17v (Typical droop from stock voltage of 1.25v @ max load)

That means the VRMs that feed the GPU on the card can only take ~300w @ 100C.

This is the reason why 290s and 290x's crash so often when they are at max load @ default clocks when you turn off all throttling if you allow the VRM to get over 80C.[/QUOT]
These calculations means absolutely nothing, just taking 1V output voltage instead of 1.2 will reduce the calculated power by 20%, when i see such blatant and willfull miscalculations i dont even bother reading the rest.

Read my entire post.

Especially the part that actually shows up in your quoting of me.

Learn to read or GTFO.

 

[Abwx]

You deny the very document that your compatriot linked?

The document say 175°C max OPERATING junction temperature, the same document specify the junction/case thermal resistance wich allow to compute the junction temperature for a given case temperature and when adding the heatsink thermal resistance to compute the temp difference between heatsink, ambiant and junction for a given TDP.

 

[3DVagabond]

You deny the very document that your compatriot linked?

That's some convenient compartmentalization you have going on there.

I've disproved your guys' crap for over 1000 posts now and I have been proven correct by time 100% of the time.

You wanna bet on you being correct?

Hell, I even got the actual TDP of the 290 and 290x and the 295x2 down to a few percentage points >>BEFORE THE CARDS WERE EVEN SOFT-LAUNCHED<<

According to your guy's information, I've beaten both the market and first party predictions for earnings trends for NVIDA and AMD/ATi for 3 quarters now.

According to your guy's information, I've absolutely trounced both AMD/ATi and Nvidia's entire competitive analysis departments for over 1 year now.

I don't have a compatriot. I have no idea what you are implying. I stand by what I've said. You can heap all of the praise on yourself that you'd like. It doesn't address anything I've said. VRM's @ ~100°C is common.

 

[24601]

[Forget] it.

Wallow in your own filth.

I'm out.

Yes, you are. We'll see you next week.
-ViRGE

 

[Abwx]

I don't have a compatriot. I have no idea what you are implying. I stand by what I've said. You can heap all of the praise on yourself that you'd like. It doesn't address anything I've said. VRM's @ ~100°C is common.

And it is happy that they are hardly hot at those temps given that they almost always run hotter then the GPUs once some power is drained , that s why i dont understand the scandal about 20°C variation in thermal designs around thoses components since everything is still fully running under specs, i guess that the GPUs and CPUs much more limited thermals are assumed as being the rule for anything silicium based.

 

[96Firebird]

It was in another of their review, IIRC the specified temp was 125°C,
i ll look for the exact article.

Ever find this?

 

[Silverforce11]

The hottest VRM operating temps I've seen were on the XFX DD 7970, 115C (no VRM heatsink, piss weak noisy fans). GTX460, 480 and a few others were running around 100C as well. Generally the accepted "safe" limit is 125C, but its not advised simply because mosfets drop in efficiency hard at above 100C.

The point is that people see VRM temps of these cards in the 80+ range and think "Oh! Thats HOT, must be bad".. except that most high end cards have VRMs in that range, including thermal imaging of 780s and 780ti reference or custom cards (per guru3d and hardware.fr), some are in the 90C range.

It is true what commie said, R290s with its 5 phase is power limited so bigger OC with high vcore is not a good mix. Even under water, it seems to peak around 1.25ghz.

But all this is irrelevant since its the Titan Z we're talking about and who knows how many phases NV decide to give it? Bit too premature to worry about it.