Nvidia Pascal Announcement

Notebook 1050 Ti Specs and Benchmarks

Base core clock 1490 MHz
Boost core clock 1624 MHz
CUDA cores 768
Memory bandwidth 112.1 GB/s
Memory type and size GDDR5, 4GB
Memory interface 128-bit
ROPs / TMUs 32 / 64

http://laptopmedia.com/highlights/nvidia-geforce-gtx-1050-ti-laptop-specs-and-benchmarks

7-10% faster than mobile GTX 970 (100W), up to 86% faster than GTX 960. Performance on par with the desktop version based on the benchmark results.

More readable for the folks using the dark theme.
 
Huum, so they have set the 1050TI in laptop at 200mhz higher than the desktop counterpart. thats a first.

Not exactly..for example GTX960M (GM107) was clocked higher than the desktop 750Ti (~80mhz higher base clock). But 200 Mhz seems a bit of a stretch..especially since the other mobile Pascals have lower base clocks than their desktop counterparts. Even if we assume the same clock as the desktop though..the performance should be close to 970M..impressive!
 
ZOTAC Announces VR GO Backpack PC with GeForce GTX 1070
ZOTAC this week has formally introduced its VR GO backpack PC designed for virtual reality enthusiasts. The system is equipped with NVIDIA’s GeForce GTX 1070 graphics, a quad-core CPU from Intel, rich I/O capabilities and a hot-swappable battery. Despite its positioning, the VR GO can be used like a normal desktop computer, which differentiates it from early prototypes of wearable VR PCs from other makers.
http://www.anandtech.com/show/10807/zotac-announces-vr-go-backpack-pc-with-geforce-gtx-1070
 

Anyone else find it interesting that they state that the combined tdp of the cpu and gpu is about 150W to cool?

If it's a mobile 45ishW CPU, then that's one more data point that suggests that the laptop 1070 puts off about 100ishW of heat.

Not earthshattering, but secret mobile TDPs make these kinds of data points helpful.
 
Not exactly..for example GTX960M (GM107) was clocked higher than the desktop 750Ti (~80mhz higher base clock). But 200 Mhz seems a bit of a stretch..especially since the other mobile Pascals have lower base clocks than their desktop counterparts. Even if we assume the same clock as the desktop though..the performance should be close to 970M..impressive!


Only one Pascal mobile has less base clocks, that is the 1070, and that is because its got more functional shader units. All other Pascals *outside of the 1050ti* have the same base clocks as their desktop variants.
 
Anyone else find it interesting that they state that the combined tdp of the cpu and gpu is about 150W to cool?

If it's a mobile 45ishW CPU, then that's one more data point that suggests that the laptop 1070 puts off about 100ishW of heat.

Not earthshattering, but secret mobile TDPs make these kinds of data points helpful.


Well this was the whole discussion about TDP, its one thing for just a card and its components or a CPU, the form factor makes a big difference because if they are able to move heat away faster due to design, the TDP will be greater due to the design.

But yeah its interesting that its only 150 watts, 45 watts for the CPU and 105 watts for the GPU seems like they did a pretty good job for an I7 (no specs so I'm assuming its configurable) and a 1070.
 
But yeah its interesting that its only 150 watts, 45 watts for the CPU and 105 watts for the GPU seems like they did a pretty good job for an I7 (no specs so I'm assuming its configurable) and a 1070.

Indeed, take your pick of quad-core mobile skylakes.

I assume mobile 1070 is just underclocked in the mid-1.4GHz range? I suppose it's nice to attach a figure to that.
 
Anyone else find it interesting that they state that the combined tdp of the cpu and gpu is about 150W to cool?

If it's a mobile 45ishW CPU, then that's one more data point that suggests that the laptop 1070 puts off about 100ishW of heat.

Not earthshattering, but secret mobile TDPs make these kinds of data points helpful.

Not very surprising since Nvidia stated that the Mobile 1070 has a similar TDP to 980M which was ~100W.
Only one Pascal mobile has less base clocks, that is the 1070, and that is because its got more functional shader units. All other Pascals *outside of the 1050ti* have the same base clocks as their desktop variants.

I'm assuming that was from your lower back again..because like I said (and Nvidia seems to agree with me here...) the base clock for mobile 1080 is 1556 Mhz vs 1607 Mhz desktop, for the 1070 it is 1442 Mhz vs 1506 Mhz Desktop and for the 1060 it is 1404 Mhz vs 1506 Mhz Desktop - http://www.geforce.com/hardware/10series/notebook
I assume mobile 1070 is just underclocked in the mid-1.4GHz range? I suppose it's nice to attach a figure to that.
Yep! Mobile 1070 has a base clock of 1442 mhz.
 
Indeed, take your pick of quad-core mobile skylakes.

I assume mobile 1070 is just underclocked in the mid-1.4GHz range? I suppose it's nice to attach a figure to that.

The desktop efficiency figures produced by Tom's Hardware can also be of interest, they did similar for the 480 and 1060.
They use Metro Last Light at 4k as it is one of the heavier power demand titles from their experience on both AMD and Nvidia.
01-Power-Consumption-vs-Gaming-Performance-FPS.png


r_600x450.png

More details can be found in the following URL, the interesting one is MSI 1070 card and how close all its results are from a frame time in ms perspective:http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1070-8gb-pascal-performance,4585-8.html
Seems the 1070 is a good candidate considering price, strong performance and efficiency requirement settings in laptops.
Their conclusion as well:
The GeForce GTX 1070 Founders Edition turns out to be extremely efficient at low power consumption levels. Compared to the stock settings, it provides more than 75 percent of its gaming performance at only 50 percent power.
But then the results for the 1060 from an efficiency perspective are pretty incredible, 62W at base clock when gaming.
It does seem this is one area Pascal has really jumped beyond Maxwell, in terms of performance at low power consumption settings.
Cheers
 
Last edited:
I'm assuming that was from your lower back again..because like I said (and Nvidia seems to agree with me here...) the base clock for mobile 1080 is 1556 Mhz vs 1607 Mhz desktop, for the 1070 it is 1442 Mhz vs 1506 Mhz Desktop and for the 1060 it is 1404 Mhz vs 1506 Mhz Desktop - http://www.geforce.com/hardware/10series/notebook

Thanks for sharing that.

I knew the notebook pascal variants had lower clocks (there's no free lunch, lol), but I had been referencing the Wikipedia page.

https://en.m.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units

Much better when it comes from the horse's mouth.
 
Thanks for sharing that.

I knew the notebook pascal variants had lower clocks (there's no free lunch, lol), but I had been referencing the Wikipedia page.

Much better when it comes from the horse's mouth.

You're welcome..and yes..aint that the truth! There's no magical way to reduce TDP unfortunately. But in most cases we'll be within 10-20% of the desktop performance..so its still quite good.
1060 in my razer blade seems to have baseclock 1.405Ghz and boost 1.671GHz

Yes that is the official spec. Does it boost beyond the official boost clock at any time? I would think not because of the lower power limits and cooling.
 
You're welcome..and yes..aint that the truth! There's no magical way to reduce TDP unfortunately. But in most cases we'll be within 10-20% of the desktop performance..so its still quite good.
Just to add.

Seems to be pretty similar battery power used to the 980M but with nearly double the performance.
Hardware Canucks compared a laptop from the same manufacturer with identical spec for 980M and the mobile variant 1070.
Models to reference are the Eurocom Tornado GTX1070 and also Eurocom Tornado GTX980M.
1070-980M-86.jpg


And quick snapshot of 2 games, known to be one good and one bad for Nvidia:

1070-980M-74.jpg


1070-980M-72.jpg


The review approach summary is:
In an effort to overcome these hurdles, we’ve teamed up with Eurocom since they offer a broad range of drop-in GPU upgrade options for many of their notebook lineups. Provided the mounts are the same and the cooling system compatible, this allows customers to bring their slightly older notebooks up to a more current set of expectations for a fraction of the cost of buying a brand new computer. This also grants us the unique opportunity to test some mobile GPUs in a direct head to head comparison on identical systems.

In this case I’m using Eurocom’s 15.6” Tornado F5 chassis which is quite a bit more compact than the ASUS monster and weighs in at 6.5lbs. What it lacks in size it more than makes up for in portability and features. Our sample, which rings in at just under $2100USD, includes an i7-6700K, 16GB of memory, a 256GB M.2 SSD and a 1080P 60Hz G-SYNC display even when equipped with a GTX 1070. All told, that’s a pretty good price considering we’ve already seen very similar notebooks run for hundreds more.

While those specs are certainly something to drool over regardless of whether you are a desktop or notebook user, this system is also compatible with a good number of GPUs due to its MXM 3.0 interface. Right now it’s compatible with the GTX 1070, GTX 1060, GTX 980M, GTX 970M along with no fewer than four different Quadro cards for the professional users out there

http://www.hardwarecanucks.com/foru...notebook-gpu-battle-gtx-1070-vs-gtx-980m.html

Cheers
 
Last edited:
Yes that is the official spec. Does it boost beyond the official boost clock at any time? I would think not because of the lower power limits and cooling.

I'm not sure how to really get proper measurement. I used gpu-z log to file and ran 3dmark time spy. The clocks look to hover around 1580Mhz +-20MHz. Peak was 1657.5MHz. Memory seems to be 2003.4MHz. Peak GPU temp was 74C. Interestingly after timespy was done gpu clocked itself way way down to 139.0Mhz. The gpu score was 3 407. I have no idea how that compares.

On desktop use while browsing and playing youtube video I see gpu 607.5MHz memory 202.5MHz gpu temp stable 49C.

The machine is clearly cooler and quieter on desktop use when disabling nvidia gpu and using the intel integrated one.
 
Last edited:
I'm not sure how to really get proper measurement. I used gpu-z log to file and ran 3dmark time spy. The clocks look to hover around 1580Mhz +-20MHz. Peak was 1657.5MHz. Memory seems to be 2003.4MHz. Peak GPU temp was 74C. Interestingly after timespy was done gpu clocked itself way way down to 139.0Mhz. The gpu score was 3 407. I have no idea how that compares.

On desktop use while browsing and playing youtube video I see gpu 607.5MHz memory 202.5MHz gpu temp stable 49C.

The machine is clearly cooler and quieter on desktop use when disabling nvidia gpu and using the intel integrated one.

That seems to be on the low side actually. A desktop 1070 has a GPU score of about 5500 with Async compute on. I would have expected the mobile version to score at least 4500. This test is being run with the laptop plugged into the mains right?
 
That seems to be on the low side actually. A desktop 1070 has a GPU score of about 5500 with Async compute on. I would have expected the mobile version to score at least 4500. This test is being run with the laptop plugged into the mains right?

it's a 1060 on razer blade. It was plugged in. I think it was running proper as the clocks were in boost range.
 
it's a 1060 on razer blade. It was plugged in. I think it was running proper as the clocks were in boost range.
Ohh right...my bad..I had just read the 1070 post above so got them mixed up! That score is absolutely fine for a 1060.

I'm just waiting for the 1050Ti and 1050 mobile to launch. Just a bit over two weeks to go for Black Friday..hoping we see some laptops by then.
 
it's a 1060 on razer blade. It was plugged in. I think it was running proper as the clocks were in boost range.
In the Nvidia driver control panel there is an option called Power management mode under Manage 3D settings. For Pascal Nvidia added a new default setting 'Optimal Power'. If you want stable clocks in the boost range without clock drops when playing games or benchmarking then you should change it to 'Prefer maximum performance' if you haven't already done it.
 
Back
Top