Nvidia Begins Seeding Reviewers With the New TITAN-X with Pascal - Synthetic CuDNN Benchmarks Show Up To 200% Performance Increase

Constructed benchmarks for the Nvidia TITAN X have leaked out on Chiphell (via Videocardz) and show a very significant performance increase over the older generation Geforce GTX TITAN X. This of course also means that Nvidia has started sampling reviewers with units (or at to the lowest degree the initial batch has begun shipping). The brand new TITAN X has 3584 CUDA cores and clocked at 1531 Mhz (Boost) offers a pregnant hardware upgrade over its older brother which boasted just 3072 CUDA cores and a measly clock speed of 1075 Mhz (Boost).

The new TITAN Ten starts shipping to select reviewers - offers anywhere from a 63% to a 200% speedup with CuDNN5

Since Nvidia has deliberately dropped the Geforce GTX branding from the new TITAN X, instead branding it as a 'prosumer' card, constructed benches that focus on CuDNN are pretty relevant for this product. The new TITAN X offers a speedup of up to 2x in some cases - which is pretty huge. Keep in heed however, that the new TITAN 10 is using a newer implementation of CuDNN and so the speedup is non simply hw gains. That said, gains from version differences are unremarkably quite small in nature so the higher clock speed besides as the increased amount of cores tin exist safely said to account for the vast majority of the speedup.

On newspaper the new Titan has 16% more than cores and a 42% higher clock rate. This means that correct off the bat (and not accounting for whatsoever improved CuDNN library optimization and or architectural gains) yous are looking at a raw compute increase of 66%. If you subtract that figure from the speedup gained, you will arrive at the true architectural gains (Maxwell to Pascal) and the gains from using CuDNN5 instead of CuDNN4. In some cases, these true gains are meaning in some cases all the same, they are actually a fleck under what we would look the blank minimum speedup to be.

According to the benchmarks y'all are looking at a speedup of betwixt 74% to 91% on Alexnet, 76% to 200% on OverFeat, 74% to 884% on Inception and 91% to 98% on VGG. This ways that in compute work that involves deep neural nets the new TITAN X will offer a significant speedup. It has better ability efficiency as well. However, when we talk well-nigh the price, the MSRP is going to exist 1200 dollars and because TITANs rarely sell at MSRP (the former $1000 MSRP Titan is retailing in the $1700-1900 range on Amazon and Newegg) the toll might give a potential client pause. Besides considering nosotros don't know anything about DP simply yet, information technology would make more sense for the prosumer market to pursue more than value oriented purchases,like lets say, multiple GTX 1070s.

NVIDIA GeForce 10 Pascal Family

Graphics Carte Name	NVIDIA GeForce GTX 1050 2 GB	NVIDIA GeForce GTX 1050 3 GB	NVIDIA GeForce GTX 1050 Ti	NVIDIA GeForce GTX 1060 3 GB	NVIDIA GeForce GTX 1060 five GB	NVIDIA GeForce GTX 1060 6 GB	NVIDIA GeForce GTX 1070	NVIDIA GeForce GTX 1070 Ti	NVIDIA GeForce GTX 1080	NVIDIA Titan X	NVIDIA GeForce GTX 1080 Ti	NVIDIA Titan Xp
Graphics Core	GP107	GP107	GP107	GP106 / GP104	GP106	GP106 / GP104	GP104	GP104	GP104	GP102	GP102	GP102
Procedure Node	14nm FinFET	14nm FinFET	14nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET	16nm FinFET
Dice Size	132mm2	132mm2	132mm2	200mm2	200mm2	200mm2	314mm2	314mm2	314mm2	471mm2	471mm2	471mm2
Transistors	3.3 Billion	three.3 Billion	3.three Billion	four.4 Billion	4.4 Billion	iv.4 Billion	seven.2 Billion	7.ii Billion	vii.2 Billion	12 Billion	12 Billion	12 Billion
CUDA Cores	640 CUDA Cores	768 CUDA Cores	768 CUDA Cores	1152 CUDA Cores	1280 CUDA Cores	1280 CUDA Cores	1920 CUDA Cores	2432 CUDA Cores	2560 CUDA Cores	3584 CUDA Cores	3584 CUDA Cores	3840 CUDA Cores
Base of operations Clock	1354 MHz	1392 MHz	1290 MHz	1506 MHz	1506 MHz	1506 MHz	1506 MHz	1607 MHz	1607 MHz	1417 MHz	1480 MHz	1480 MHz
Boost Clock	1455 MHz	1518 MHz	1392 MHz	1708 MHz	1708 MHz	1708 MHz	1683 MHz	1683 MHz	1733 MHz	1530 MHz	1583 MHz	1582
FP32 Compute	ane.8 TFLOPs	two,3 TFLOPs	2.1 TFLOPs	iv.0 TFLOPs	4.4 TFLOPs	4.iv TFLOPs	6.5 TFLOPs	8.1 TFLOPs	ix.0 TFLOPs	xi TFLOPs	11.5 TFLOPs	12.five TFLOPs
VRAM	2 GB GDDR5	3 GB GDDR5	iv GB GDDR5	3 GB GDDR5	6 GB GDDR5	six GB GDDR5/X	8 GB GDDR5/X	8 GB GDDR5	eight GB GDDR5X	12 GB GDDR5X	11 GB GDDR5X	12 GB GDDR5X
Memory Speed	7 Gbps	7 Gbps	7 Gbps	8 Gbps	8 Gbps	9 Gbps / 10 Gbps	viii Gbps	8 Gbps	11 Gbps	10 Gbps	11 Gbps	11.4 Gbps
Retentiveness Bandwidth	112 GB/southward	84 GB/s	112 GB/due south	192 GB/s	160 GB/s	224 GB/s / 240 GB/south	256 GB/due south	256 GB/southward	352 GB/due south	480 GB/south	484 GB/s	547 GB/south
Coach Interface	128-chip bus	96-bit bus	128-bit bus	192-bit bus	160-bit motorcoach	192-bit bus	256-bit bus	256-bit autobus	256-bit motorcoach	384-bit bus	352-chip bus	384-bit jitney
Power Connector	None	None	None	Unmarried half-dozen-Pin Power	Single 6-Pin Power	Single vi-Pin Ability	Single 8-Pivot Power	Single 8-Pin Power	Single 8-Pin Power	8+half-dozen Pin Power	8+6 Pin Ability	8+6 Pin Power
TDP	75W	75W	75W	120W	120W	120W	150W	180W	180W	250W	250W	250W
Display Outputs	1x Brandish Port 1.four 1x HDMI 2.0b 1x DVI	1x Display Port 1.iv 1x HDMI ii.0b 1x DVI	1x Display Port 1.4 1x HDMI two.0b 1x DVI	3x Display Port 1.4 1x HDMI two.0b 1x DVI	3x Display Port i.iv 1x HDMI ii.0b 1x DVI	3x Display Port 1.4 1x HDMI ii.0b 1x DVI	3x Display Port 1.4 1x HDMI 2.0b 1x DVI	3x Display Port i.4 1x HDMI 2.0b 1x DVI	3x Brandish Port 1.4 1x HDMI 2.0b 1x DVI	3x Brandish Port ane.4 1x HDMI two.0b 1x DVI	3x Brandish Port 1.four 1x HDMI 2.0b	3x Display Port 1.iv 1x HDMI 2.0b
Launch Date	October 2022	May 2022	Oct 2022	September 2022	August 2022	July 2022	June 2022	October 2022	May 2022	August 2022	March 2022	Apr 2022
Launch Price	$109 United states	$119 Us-$129 Usa	$139 US	$199 United states	TBD	$249 US	$349 Us	$449 US	$499 US	$1200 US	$699 US	$1200 United states