Nvidia Titan

Nvidia Titan is a series of desktop - graphics processors of the company Nvidia . The Titan graphics cards were initially assigned to the GeForce series and generally represented the most powerful model in the current series. As a special feature, some of the models had properties that are normally only found in the professional series, Quadro and Tesla (such as increased performance when calculating with double precision ). When the Titan X came onto the market in August 2016, Nvidia for the first time renounced the name GeForce and the classification in the GeForce 10 series at the time . With the Titan V at the end of 2017, an architecture was used for the first time (Volta architecture on the GV100) that is not used in the normal GeForce series. The Titan series now functions as a hybrid series between the GeForce and professional series, Quadro and Tesla.

Data overview

Graphics processors

Graphics chip	production			units							L2 cache	API support					Video processor	Bus interface stelle
	production process	transis- interfere	The - area	ROP partitions	ROPs	Unified shaders		Texture units	Tensor cores	RT cores		DirectX	OpenGL	OpenCL	CUDA	volcano
	production process	transis- interfere	The - area	ROP partitions	ROPs	Stream processors	Shader - cluster	Texture units	Tensor cores	RT cores		DirectX	OpenGL	OpenCL	CUDA	volcano
GK110 (A1 / B1)	28 nm	7.08 billion	561/533 mm²	6th	48	2880	15th	240	-	-	1536 kB	11	4.6	1.2	3.5	1.1.95	VP5	PCIe 3.0
GM200	28 nm	8 billion	601 mm²	6th	96	3072	24	192	-	-	3072 kB	12.1			5.2		VP6
GP102	16 nm	12.0 billion	471 mm²	12	96	3840	30th	240	-	-	3072 kB				6.1		k. A.
GV100	12 nm	21.1 billion	815 mm²	6th	128	5376	42	336	640	-	6144 kB			2.0	7.0		k. A.
TU102	12 nm	18.6 billion	754 mm²	6th	96	4608	72	288	576	72	6144 kB			2.0	7.5		k. A.

Model data

model	Official launch	Graphics processor (GPU)									Graphics memory			Performance data					Power consumption
		Type	Active units						Chip clock		Memory size	Storage cycle	Storage interface	Computing power (in GFlops )		Fill rate		Memory bandwidth (in GByte / s)	MGCP	Readings
		Type	ROPs	Shader - cluster	ALUs	Texture units	Tensor cores	RT cores	default	Boost	Memory size	Storage cycle	Storage interface	SP (MAD)	DP (FMA)	Pixels (in GP / s)	Texel (in GT / s)	Memory bandwidth (in GByte / s)	MGCP	Idle	3D load
GeForce GTX Titan	19th Feb. 2013	GK110-400-A1	48	14th	2688	224	-	-	837 MHz	876 MHz	6 GB GDDR5	1502 (3004) MHz	384 bits	4499.7	1311.7	33.5	187.5	288.4	250 W	12 W.	206 W
GeForce GTX Titan Black	Feb. 18, 2014	GK110-430-B1	48	15th	2880	240	-	-	889 MHz	980 MHz	6 GB GDDR5	1753 (3506) MHz	384 bits	5120.6	1706.9	35.6	213.4	336.6	250 W	13 W.	230 W
GeForce GTX Titan Z	May 28, 2014	2 × GK110-350-B1	2 × 48	2 × 15	2 × 2880	2 × 240	-	-	705 MHz	876 MHz	2 × 6 GB GDDR5	1753 (3506) MHz	2 × 384 bits	2 x 4060.8	2 x 1353.6	2 x 28.2	2 x 169.2	2 x 336.6	375 W	21 W	365 W
GeForce GTX Titan X	March 17, 2015	GM200-400-A1	96	24	3072	192	-	-	1002 MHz	1075 MHz	12 GB GDDR5	1753 (3506) MHz	384 bits	6604	206	103.2	206.4	336.6	250 W	13 W.	240 W
Titan X	Aug 2, 2016	GP102-400-A1	96	28	3584	224	-	-	1417 MHz	1531 MHz	12 GB GDDR5X	2500 (5000) MHz	384 bits	10974	343	147	342.9	480	250 W	11 W.	220 W
Titan Xp	Apr 6, 2017	GP102-450-A1	96	30th	3840	240	-	-	1404 MHz	1582 MHz	12 GB GDDR5X	2850 (5700) MHz	384 bits	12150	380	151.9	379.7	547.2	250 W	12 W.	245 W
Titan V	7th Dec 2017	GV100-400-A1	96	80	5120	320	640	-	1200 MHz	1455 MHz	12 GB HBM2	425 (850) MHz	3072 bits	14899	7450	139.7	465.6	652.8	250 W	k. A.	k. A.
Titan V "CEO Edition"	Jun 21, 2018	GV100-450-A1	128	80	5120	320	640	-	1200 MHz	1530 MHz	32 GB HBM2	425 (850) MHz	4096 bits	15667	7834	195.8	489.6	870.4	250 W	k. A.	k. A.
Titan RTX	Dec 18, 2018	TU102-400A-A1	96	72	4608	288	576	72	1350 MHz	1770 MHz	24 GB GDDR6	3500 (7000) MHz	384 bits	16312	510	169.9	509.8	672	280 W	k. A.	k. A.

Remarks

↑ The date indicated is the date of the public presentation, not the date of availability of the models.
↑ The specified performance values for the computing power via the stream processors, the pixel and texel fill rate, as well as the memory bandwidth are theoretical maximum values (with boost clock) that are not directly comparable with the performance values of other architectures. The overall performance of a graphics card depends, among other things, on how well the available resources can be used or fully utilized. There are also other factors that are not listed here that affect performance.
↑ ^a ^b The specified clock rates are the reference data recommended or specified by Nvidia; the I / O clock is specified for the memory clock (the effective clock in brackets). However, the exact clock rate can deviate by a few megahertz due to different clock generators, and the final definition of the clock rates is in the hands of the respective graphics card manufacturer. It is therefore entirely possible that there are or will be graphics card models that have different clock rates.
↑ The MGCP value specified by Nvidia does not necessarily correspond to the maximum power consumption. This value is also not necessarily comparable with the TDP value of the competitor AMD.
↑ The measured values listed in the table relate to the pure power consumption of graphics cards that correspond to the Nvidia reference design. A special measuring device is required to measure these values; Depending on the measurement technology used and the given measurement conditions, including the program used to generate the 3D load, the values can fluctuate between different devices. Therefore, measured value ranges are given here, each representing the lowest, typical and highest measured values from different sources.

Individual evidence

↑ NVIDIA GeForce GTX TITAN. Nvidia Corporation, accessed June 4, 2013 .
↑ ^a ^b Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .
↑ NVIDIA GeForce GTX TITAN Black. Nvidia Corporation, accessed February 18, 2014 .
↑ ^a ^b ^c ^d Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .
↑ NVIDIA GeForce GTX TITAN Z. Nvidia Corporation, accessed May 28, 2014 .
↑ NVIDIA GeForce GTX Titan X. Nvidia Corporation, accessed March 17, 2015 .
↑ ^a ^b Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .
↑ NVIDIA TITAN X. Nvidia Corporation, July 21, 2016, accessed July 22, 2016 .
↑ ^a ^b ^c ^d Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .
↑ NVIDIA TITAN Xp. Nvidia Corporation, April 6, 2017, accessed April 6, 2017 .
↑ ^a ^b NVIDIA TITAN V. Nvidia Corporation, accessed September 13, 2018 .
↑ NVIDIA TITAN RTX. Nvidia Corporation, December 18, 2018, accessed January 3, 2019 .

[Launch-1] The date indicated is the date of the public presentation, not the date of availability of the models.

[Leistung-2] The specified performance values for the computing power via the stream processors, the pixel and texel fill rate, as well as the memory bandwidth are theoretical maximum values (with boost clock) that are not directly comparable with the performance values of other architectures. The overall performance of a graphics card depends, among other things, on how well the available resources can be used or fully utilized. There are also other factors that are not listed here that affect performance.

[Takt-3] The specified clock rates are the reference data recommended or specified by Nvidia; the I / O clock is specified for the memory clock (the effective clock in brackets). However, the exact clock rate can deviate by a few megahertz due to different clock generators, and the final definition of the clock rates is in the hands of the respective graphics card manufacturer. It is therefore entirely possible that there are or will be graphics card models that have different clock rates.

[MGCP-4] The MGCP value specified by Nvidia does not necessarily correspond to the maximum power consumption. This value is also not necessarily comparable with the TDP value of the competitor AMD.

[5] The measured values listed in the table relate to the pure power consumption of graphics cards that correspond to the Nvidia reference design. A special measuring device is required to measure these values; Depending on the measurement technology used and the given measurement conditions, including the program used to generate the 3D load, the values can fluctuate between different devices. Therefore, measured value ranges are given here, each representing the lowest, typical and highest measured values from different sources.

[6] NVIDIA GeForce GTX TITAN. Nvidia Corporation, accessed June 4, 2013 .

[3dc_1-7] Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .

[8] NVIDIA GeForce GTX TITAN Black. Nvidia Corporation, accessed February 18, 2014 .

[3dc_2-9] Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .

[10] NVIDIA GeForce GTX TITAN Z. Nvidia Corporation, accessed May 28, 2014 .

[11] NVIDIA GeForce GTX Titan X. Nvidia Corporation, accessed March 17, 2015 .

[3dc_3-12] Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .

[13] NVIDIA TITAN X. Nvidia Corporation, July 21, 2016, accessed July 22, 2016 .

[3dc_4-14] Power consumption of current and past graphics cards. 3DCenter.org, accessed September 13, 2018 .

[15] NVIDIA TITAN Xp. Nvidia Corporation, April 6, 2017, accessed April 6, 2017 .

[Titan_V-16] NVIDIA TITAN V. Nvidia Corporation, accessed September 13, 2018 .

[17] NVIDIA TITAN RTX. Nvidia Corporation, December 18, 2018, accessed January 3, 2019 .