2016 nMP

[netkas]

Anyone wants overclocked cards benchmarks ?

http://www.overclockersclub.com/reviews/nvidia_geforcegtx_1080_overclocking/

owners of 980ti/titanx and 4k display will probably want to overclock the card and wait for next big chip, rest may want to upgrade.

 

[AidenShaw]

Any CUDA benchmarks? That's far more interesting given OpenCL's fading away.

Especially interested to see FP16 benchmarks....

 

[linuxcooldude]

This image explains why not everyone happy about 1080

Spoiler: Pic

While I don't think performance per dollar is always applicable in a professional environment, just like I consider how under-clocked cards a moot point in nMP since the performance output is greater than a Titan X. Price per dollar is more important perhaps in the consumer relm where what they do is not in a paid profession that justifies the cost and pays for itself in a set amount of time.

Many may not use the 1080 in a workstation and may be waiting for the Quattro version thats coming at a much higher price anyway.

 

[beaker7]

Many may not use the 1080 in a workstation and may be waiting for the Quattro version thats coming at a much higher price anyway.

Waiting semi patiently here...

 

[tuxon86]

While I don't think performance per dollar is always applicable in a professional environment, just like I consider how under-clocked cards a moot point in nMP since the performance output is greater than a Titan X. Price per dollar is more important perhaps in the consumer relm where what they do is not in a paid profession that justifies the cost and pays for itself in a set amount of time.

Many may not use the 1080 in a workstation and may be waiting for the Quattro version thats coming at a much higher price anyway.

In what? FCPX and LogicX?

 

[linuxcooldude]

In what? FCPX and LogicX?

Any third party software can utilize dual GPU's on the nMP. All they need to do is use Xcode. Apple even has a new programming language called swift to make it even easier. If your software of choice does not support dual GPU's needs to contact said software developer and ask for it.

 

[viersen]

Does anybody here knows if Apple is going to release this June a new update for the MP (MP7.1)???
Any positive and serious rumors regarding the WWDC?

 

[ManuelGomes]

All is quiet from Apple's side, no news whatsoever unfortunately.
We can only wish and speculate at the moment.
Most worthy updates are already in place, or close enough.
Still, any guess is a good guess right now.

Where is that info on the new PCI IDs from AMD Polaris cards? Any info on the specs?
[doublepost=1463595461][/doublepost]GP104 seems to have 1:32 DP ratio, according to the info above. Is this correct?
[doublepost=1463595872][/doublepost]Driving 8K displays is awesome for Pascal.
Still, I can't find info on HDMI 2.0b, at least not on wiki.
koyoot, you released some info but I can only find 2.0, 2.0a and 2.1, not 2.0b supported by Pascal. What does it bring new to the table?
NVidia going HDR (10b/12b) finally is a good sign.

 

[tuxon86]

Any third party software can utilize dual GPU's on the nMP. All they need to do is use Xcode. Apple even has a new programming language called swift to make it even easier. If your software of choice does not support dual GPU's needs to contact said software developer and ask for it.

In other word only FCPX and Logic X.

 

[linuxcooldude]

In other word only FCPX and Logic X.

Nope. If a small developer, such as from Pixelmator, can use the full power of the nMP almost from the initial release of the nMP in 2013, any developer can do it. DaVinci resolve, Premiere Pro ( But needs better support ) come to mind too.

 

[koyoot]

All is quiet from Apple's side, no news whatsoever unfortunately.
We can only wish and speculate at the moment.
Most worthy updates are already in place, or close enough.
Still, any guess is a good guess right now.

Where is that info on the new PCI IDs from AMD Polaris cards? Any info on the specs?
[doublepost=1463595461][/doublepost]GP104 seems to have 1:32 DP ratio, according to the info above. Is this correct?
[doublepost=1463595872][/doublepost]Driving 8K displays is awesome for Pascal.
Still, I can't find info on HDMI 2.0b, at least not on wiki.
koyoot, you released some info but I can only find 2.0, 2.0a and 2.1, not 2.0b supported by Pascal. What does it bring new to the table?
NVidia going HDR (10b/12b) finally is a good sign.

Here you are

 

[Larry-K]

Any third party software can utilize dual GPU's on the nMP. All they need to do is use Xcode. Apple even has a new programming language called swift to make it even easier. If your software of choice does not support dual GPU's needs to contact said software developer and ask for it.

They could, but apparently, not many want to.

I postulate that Apple is just too small of a workstation (okay, it's a stretch) vendor, and has shown too little interest in the MacPro for people to want to put the effort out.

But, I'm sure a quick phone call to my software vendors will get them to immediately throw resources behind said request.

 

[linuxcooldude]

They could, but apparently, not many want to.

I postulate that Apple is just too small of a workstation (okay, it's a stretch) vendor, and has shown too little interest in the MacPro for people to want to put the effort out.

But, I'm sure a quick phone call to my software vendors will get them to immediately throw resources behind said request.

Could be a lot of reasons: Too lazy, not feasible, its too hard, not hiring the right programmers, cost too much. The resources are there, ( GPU's) but the programmers need to take advantage of it to be more useful to everybody.

But, I'm sure a quick phone call to my software vendors will get them to immediately throw resources behind said request.

But I'm sure coming on an internet forum to complain about it would get it done faster.

 

[tuxon86]

Nope. If a small developer, such as from Pixelmator, can use the full power of the nMP almost from the initial release of the nMP in 2013, any developer can do it. DaVinci resolve, Premiere Pro ( But needs better support ) come to mind too.

So again, only FCPX and Logic really. OpenCL is dying on the vine and so is the nMP.

 

[MacVidCards]

Anyone wants overclocked cards benchmarks ?

http://www.overclockersclub.com/reviews/nvidia_geforcegtx_1080_overclocking/

Wowza !

1080 makes everything else look like running in place.

Most amazing graph is the last one, power draw at load while overclocking. 1080 had LOWEST POWER DRAW OF ALL, lower even than GTX970.

And this was one graph where AMD did get tallest bar, the Hawaii card managed to draw an additional 120 Watts, create a lot more heat, and still be significantly slower.

So, card that draws minimal power, creates little heat, and is still fastest on the planet. Trying to think of a computer maker who needs cards like that for their anorexic machines...

Or will they go with the AMD "Space Heater 3" line to save a few bucks and add to their stashed offshore billions? Anyone want to lay odds?

(I suppose they could always wait for the rumored "Paper StarTiger" line that threatens to reshuffle the middle-of-the-road cards. Shooting for the middle is a recent Apple trend)

 

[ManuelGomes]

Thanx. You posted it before but couldn't find it.
Strange that couldn't find any info on it.
Not capable of HDR 5K@60Hz, or 4K for that matter (only level A, that Polaris might do).

 

[MacVidCards]

So again, only FCPX and Logic really. OpenCL is dying on the vine and so is the nMP.

Rude ! Wedding Videographer/caterers across the country swear by FCPX ! Same with the YouTube cat video crowd.

(You don't know flashy until you see some of the starburst transitions you can use between shots)

 

[linuxcooldude]

So again, only FCPX and Logic really. OpenCL is dying on the vine and so is the nMP.

No, not really. Not everyone wants to be locked into CUDA. OpenCL is just a stepping stone to the next API such as Metal. OpenCL and its next or new API will still be applicable to the Linux platform. So it will continue on to its next form.

 

[ManuelGomes]

The space heater joke makes you look like an a$$, wasn't funny from the get go anyway.
As all other remarks actually.
Hoe much is NVidia paying you exactly?

 

[Roykor]

It is true. That's why it'd be Mac Pro SE. Like iPhone SE. And it should be cheaper too.

• Xeon v5
• 8GB DDR4
• Radeon 480, Fury Nano and Nvidia GTX 1080 as CTO options
• Starting price US $1999

Lets do a headless machine:

• i7 6800K 6-core / 12 treads (or the two on top of this one)
• 16 / 32 / 64 GB DDR4
• GTX 970 / 980ti / 1070 / 1080
• dual M2 SSD 512 in raid0
• + 4 extra SSD slots in the case

I will be very interested. This will be my next PC soon.

 

[linuxcooldude]

Rude ! Wedding Videographer/caterers across the country swear by FCPX ! Same with the YouTube cat video crowd.

(You don't know flashy until you see some of the starburst transitions you can use between shots)

The very same Youtube, the multimillion dollar platform that has sold a private channel to Disney for $500 million dollars?!?

 

[Roykor]

No, not really. Not everyone wants to be locked into CUDA. OpenCL is just a stepping stone to the next API such as Metal. OpenCL and its next or new API will still be applicable to the Linux platform. So it will continue on to its next form.

I dont understand these kind of answers. Lets say you need to do your job. Your tools are openCL and Cuda coded. So you have both in your software that speed up your proces. Some stuff are on opencl, and some on Cuda (like a lot in Adobe) So if a Nvidia is the best way to go to do your job faster.. than a Nvidia it is. if you dont use any cuda, you would not give a dime about a nvidia card. But when you are using the argument that you dont like to get locked in.. only because the cuda part... mehh..

 

[tuxon86]

No, not really. Not everyone wants to be locked into CUDA. OpenCL is just a stepping stone to the next API such as Metal. OpenCL and its next or new API will still be applicable to the Linux platform. So it will continue on to its next form.

OpenCL wasn't planned as a stepping stone, it just failed. Cuda is also available on Linux and doing quite well, better than any AMD pushed api of late.
[doublepost=1463599853][/doublepost]

The very same Youtube, the multimillion dollar platform that has sold a private channel to Disney for $500 million dollars?!?

Of which the use of FCPX has zilch relevancy to the value of YouTube...

 

[flat five]

In other word only FCPX and Logic X.

http://www.indigorenderer.com/forum/viewtopic.php?f=1&t=13927

 

[MacVidCards]

The very same Youtube, the multimillion dollar platform that has sold a private channel to Disney for $500 million dollars?!?

That's the one. If you are referring to "Maker" a guy who used to build sets for me had a revenue share deal on that sale. He's retired now.

I designed a booth for a YouTube convention, met Rainbowman, etc. our booth was a hit because we tossed a wet trout into people's hands with a ghost cam to record as "hilarity ensued". Nicest bunch of pimply 14 year olds I ever met.

If that is your professional market, this isn't right board, in my opinion.