Nvidia GeForce 700 series

The GeForce GTX Titan was initially assigned to the GeForce 600 series and has been part of the GeForce 700 series since the GeForce GTX 780 was introduced.

The GeForce 700 series is a series of desktop - graphics chip company Nvidia and successor of the GeForce 600 series . All graphics processors of this series support the Shader model  5.0 (SM 5.0), DirectX  11, as well as OpenCL , CUDA and thus also PhysX .

description

GK110

With 7.1 billion transistors on around 561 mm² (533 mm² in production-optimized B1 stepping), the GK110-GPU is the largest and most complex graphics processor of the Kepler generation. It consists of 2880 shader and 240 texture units, which are distributed over 15 SMX blocks (shader clusters). These, in turn, are distributed over five graphics processing clusters, which means that the GK110 has a ratio of 3: 1 (in contrast to the rest of the Kepler generation of graphics processors, which use a ratio of 2: 1). Another special feature of the GK110 is the additional 64 separate ALUs per SMX block, which are not responsible for the single-precision (FP32) but for the double-precision operation (FP64) . The features "Dynamic Parallelism", "Hyper-Q" and "GPUDirect" are also intended for the professional sector and are only available on the GK110-GPU.

On February 19, 2013, Nvidia presented the GeForce GTX Titan as a special series, using the GK110-400-A1 with 14 of 15 active shader clusters. Compared to the GeForce GTX 680 or the Radeon HD 7970 "GHz Edition" , the card showed a 25% to 40% higher performance, depending on the setting, making it the fastest single GPU card on the market. Nvidia also released numerous features that are normally only available to the professional sector (including the full FP64 performance ). Nvidia expanded the GPU Boost function (referred to as "GPU Boost 2.0") so that the chip temperature is also taken into account as a factor. In combination with a new cooling design, this enabled very low noise levels to be achieved. This was also favored by the high energy efficiency of the GK110 GPU. The GeForce GTX Titan received very good ratings in the trade press for computing power, noise development and energy efficiency, but Nvidia's high list price of US $ 999, which was sometimes even higher than that of the faster dual-GPU graphics cards GeForce GTX 690 and Radeon HD 7990 , at a price-performance ratio classified as poor. Nevertheless, the Titan was successful on the market.

GeForce GTX 780 without cooler

After Nvidia had achieved a performance patt for the Radeon HD 7970 with the GeForce GTX 680 based on the GK104 graphics processor in spring 2012, AMD responded in the summer of 2012 with a special GHz edition, which was then the fastest single GPU graphics card on the market. The GK104 of the GTX 680, which as a performance chip was not originally intended for the high-end market and no longer had sufficient reserves to counter the GHz edition, Nvidia was only able to use the GeForce GTX Titan, the fastest single GPU graphics card present on the market. However, as this special series had a focus outside the mass retail market and was not a competitor to the Radeon HD 7970 GHz Edition due to the significantly poorer price-performance ratio, Nvidia presented the GeForce GTX 780 on May 23, 2013 . With this, the GeForce 700 series was officially introduced and the GK110 graphics processor was used for a normal GeForce card for the first time . With the GK110-300-A1, three of the 15 shader clusters are deactivated and the FP64 performance artificially reduced to 1/8, which is of no relevance for 3D applications. Furthermore, Nvidia reduced the video memory compared to the Titan from 6144 to 3072 MB. Ultimately, the changes reduced the performance compared to the Titan by 7 to 8%, but this was still enough to outperform the Radeon HD 7970 GHz Edition by 15% to 30% depending on the setting. Since Nvidia reduced the list price by a third compared to the Titan, the GeForce GTX 780 had a significantly better price-performance ratio. Since the noise level and energy consumption remained roughly the same, the card received an overall positive rating in the trade press. Nvidia has been using the production-optimized GK110-300-B1 version since September 2013 without making any further hardware changes.

On November 7, 2013, Nvidia presented the GeForce GTX 780 Ti and responded to the Radeon R9 290X from AMD in order to have the fastest single GPU graphics card on the market again after it had a higher performance than the GeForce GTX Titan. For this purpose, the GK110 GPU was used for the first time in full expansion on the retail market (previously this was only the case with the Tesla K40 ). In combination with higher clock rates, the GeForce GTX 780 Ti showed up to 10% higher performance compared to the Titan. While the noise development of the GeForce GTX 780 Ti reached the level of the other GK110 cards, for which the chip temperature limit of the GPU boost function was increased from 80 ° to 83 °, it failed to achieve energy efficiency, which was probably due to the significantly increased VRAM clock is due. Since this somewhat poorer energy efficiency was still well above that of the AMD competition and is of little importance in the high-end area, the card was rated positively in the trade press due to its high performance.

On February 18, 2014 Nvidia presented the GeForce GTX Titan Black . After Nvidia assigned the GeForce GTX 780 Ti, the fastest single-GPU graphics card on the market at the time, to the normal GeForce 700 series, the Titan was no longer relevant to the market. Therefore, Nvidia launched the Titan Black, which compared to the 780 Ti had a slight increase in the clock rate, 6 instead of 3 GB VRAM and the full FP64 performance, i.e. computing with double precision , which is very important for scientific applications . However, the Titan Black could not achieve a relevant performance advantage over the GeForce GTX 780 Ti in conventional 3D applications, which is why the card was rated critically in the trade press, as Nvidia charged a price premium of around 50%. Ultimately, the Titan Black could no longer achieve the special position of the Titan, as the normal series now had alternative products in the same performance range and only acted as a transitional solution to the professional Quadro graphics cards .

On March 25, 2014, Nvidia announced the GeForce GTX Titan Z dual-GPU graphics card for the beginning of April, but after several postponements it did not actually go on sale until May 28. The card uses two fully equipped GK110 GPUs and a total of 12 GB of VRAM, but the clock rates had to be reduced significantly in order not to exceed a TDP of 375 watts.

GK104

On May 30, 2013 Nvidia presented the GeForce GTX 770 based on the GK104 graphics processor in full configuration, which was already used in the GeForce 600 series. The GeForce GTX 770 ultimately represented a refresh of the GeForce GTX 680 with slightly increased clock rates, with which this 6 to 8% higher performance reached the performance of the Radeon HD 7970 "GHz Edition" . Since Nvidia lowered the list price from 499 to 399 US- $ at the same time, a better price-performance ratio could be achieved, which, despite the naming of the Refresh, led to positive reviews in the trade press. Furthermore, the GPU boost function has been extended to version 2.0. Although Nvidia officially increased the TDP from 195 to 230 watts, the consumption values ​​remained the same. Since Nvidia uses the Titan cooler on the GeForce GTX 770, the noise level is very low, while the power consumption remains the same, which the trade press rated positively.

GeForce GTX 760 in Nvidia reference design

For the GeForce GTX 760 , which Nvidia presented on June 25, 2013, Nvidia uses the GK104 with two deactivated shader clusters. Nvidia placed the card between the GeForce GTX 660 Ti and GTX 670, which also replaced the GeForce GTX 760, and achieved roughly the performance of the Radeon HD 7950 "Boost Edition" from the competitor AMD , but with a list price of 249 US- $ a slightly better value for money. The model was rated critically in the trade press because the GeForce GTX 760 had somewhat poorer power consumption and noise values ​​compared to its two predecessors, despite the low chip voltage and the extended GPU boost function. Nvidia introduced two other versions of the GeForce GTX 760, the GeForce GTX 760 192-bit and the GeForce GTX 760 Ti. Both models were new editions of the GeForce GTX 660 (OEM) and the GTX 670, which are reserved for the OEM market .

GM107

On February 18, 2014 Nvidia presented the GeForce GTX 750 and GeForce GTX 750 Ti . The cards use the GM107 graphics processor, which is based for the first time on the first-generation Maxwell architecture and, according to Nvidia, should achieve significantly higher energy efficiency. In independent test reports, the GeForce GTX 750 Ti achieved roughly the performance of the Radeon R7 260X with a 30% lower power consumption. Despite this low power consumption, the cards received some critical reviews in the trade press, as Nvidia set the list price at US $ 119 and US $ 149 too high compared to the AMD competition and thus achieved a poor price-performance ratio.

Data overview

Graphics processors

Graphics chip	production			units						L2 cache	API support			Video pro- cessor	Bus interface stelle
	production process	transis- interfere	The - area	ROP particle functions	ROPs	Unified shaders		Texture units			DirectX	OpenGL	OpenCL
						Stream processors	Shader - cluster	TAUs	TMUs
GF108	40 nm	0.58 billion	114 mm²	1	04th	0096	02	016	016	k. A.	11.0	4.4	1.1	VP4	PCIe 2.0
GK104	28 nm	3.54 billion	294 mm²	4th	32	1536	08th	128	128	0.5 MB		4.5	1.2	VP5	PCIe 3.0
GK107		1.30 billion	118 mm²	2	16	0384	02	032	032	0.25 MB
GK110 (A1 / B1)		7.08 billion	561/533 mm²	6th	48	2880	15th	240	240	1.5 MB
GK208		k. A.	087 mm²	1	08th	0384	02	016	016	0.5 MB					PCIe 2.0
GM107		1.87 billion	148 mm²	2	16	0640	05	040	040	2 MB	12.0	4.5		VP6	PCIe 3.0

Model data

model	Official launch	Graphics processor (GPU)							Graphics memory				Performance data
		Type	Active units				Chip clock (in MHz)		size	Clock rate (in MHz)	Type	Storage interface	Computing power (in GFlops )		Polygon throughput (in million triangles / s)	Pixel fill rate (in GPixel / s)	Texel fill rate (in GTexel / s)	Memory bandwidth (in GB / s)
			ROPs	Shader - cluster	ALUs	Texture units	default	Boost					SP (MAD)	DP (FMA)
GeForce GT 710	Jan. 29, 2016	GK208	08th	01	0192	016	954	-	1 GB	891	DDR3	64 bit	366	k. A.	954	7.6	15.2	14.4
									2 GB	1252 (2504)	GDDR5							40
GeForce GT 720	13 Aug 2014	GK208	08th	01	0192	016	797	-	1 GB	891	DDR3	64 bit	306	k. A.	797	6.4	12.8	14.4
									2 GB	1252 (2504)	GDDR5							40
GeForce GT 730 (64-bit DDR3)	Jun 18, 2014	GK208	08th	02	0384	016	902	-	2 GB	891	DDR3	64 bit	692.7	28.9	902	7.2	14.4	14.4
GeForce GT 730 (128-bit DDR3)	Jun 18, 2014	GF108	04th	02	0096	016	700	-	1 GB	891	DDR3	128 bit	268.8	22.4	350	2.8	11.2	28.8
GeForce GT 730 (GDDR5)	Jun 18, 2014	GK208	08th	02	0384	016	902	-	1 GB	1252 (2504)	GDDR5	64 bit	692.7	28.9	902	7.2	14.4	40
GeForce GT 740	May 29, 2014	GK107	16	02	0384	032	993	-	2 GB	891	DDR3	128 bit	762.6	31.8	993	7.9	31.8	28.5
									1 GB	1252 (2504)	GDDR5							80.1
GeForce GTX 745	Feb. 18, 2014	GM107	16	03	0384	024	1033	-	4 GB	891	DDR3	128 bit	793.3	24.8	1033	16.5	24.8	28.8
GeForce GTX 750	Feb. 18, 2014	GM107	16	04th	0512	032	1020	1085	1 GB	1250 (2500)	GDDR5	128 bit	1044.5	32.6	1020	16.3	32.6	80
GeForce GTX 750 Ti	Feb. 18, 2014	GM107	16	05	0640	040	1020	1085	2 GB	1350 (2700)	GDDR5	128 bit	1305.6	40.8	1020	16.3	40.8	86.4
GeForce GTX 760 192-bit	Oct 21, 2013	GK104	24	06th	1152	096	823	888	1.5 GB	1452 (2904)	GDDR5	192 bits	1896.2	79	2469	19.8	79	134
									3 GB
GeForce GTX 760	Jun 25, 2013	GK104	32	06th	1152	096	980	1033	2 GB	1502 (3004)	GDDR5	256 bit	2257.9	94.1	2940	23.5-31.4	94.1	192.3
GeForce GTX 760 Ti	Oct 21, 2013	GK104	32	07th	1344	112	915	980	2 GB	1502 (3004)	GDDR5	256 bit	2459.5	102.5	3203	29.3	102.5	192.3
GeForce GTX 770	May 30, 2013	GK104	32	08th	1536	128	1046	1085	2 GB	1753 (3506)	GDDR5	256 bit	3213.3	133.9	4184	33.5	133.9	224.4
GeForce GTX 780	May 23, 2013	GK110	48	12	2304	192	863	902	3 GB	1502 (3004)	GDDR5	384 bits	3976.7	165.7	5178	27.6-34.5	165.7	288.4
	Sep 10 2013	GK110B
GeForce GTX 780 Ti	Nov 7, 2013	GK110B	48	15th	2880	240	876	928	3 GB	1753 (3506)	GDDR5	384 bits	5045.8	210.2	6570	35	210.2	336.6
GeForce GTX Titan	19th Feb. 2013	GK110	48	14th	2688	224	837	876	6 GB	1502 (3004)	GDDR5	384 bits	4499.7	1311.7	5859	33.5	187.5	288.4
GeForce GTX Titan Black	Feb. 18, 2014	GK110B	48	15th	2880	240	889	980	6 GB	1753 (3506)	GDDR5	384 bits	5120.6	1706.9	6667.5	35.6	213.4	336.6
GeForce GTX Titan Z	May 28, 2014	2 × GK110B	2 × 48	2 × 15	2 × 2880	2 × 240	705	876	2 × 6 GB	1753 (3506)	GDDR5	2 × 384 bits	2 x 4060.8	2 x 1353.6	2 x 5287.5	2 x 28.2	2 x 169.2	2 x 336.6

Power consumption data

model	Type	Consumption ( watt )				Additional power plug
		MGCP	Readings
			Idle	3D load	Maximum load
GeForce GT 710	GK208	019th	k. A.	k. A.	k. A.	no
GeForce GT 720	GK208	023	k. A.	k. A.	k. A.	no
GeForce GT 730 (64-bit DDR3)	GK208	023	k. A.	k. A.	k. A.	no
GeForce GT 730 (128-bit DDR3)	GF108	049	k. A.	k. A.	k. A.	no
GeForce GT 730 (GDDR5)	GK208	038	k. A.	k. A.	k. A.	no
GeForce GT 740	GK107	064	k. A.	k. A.	k. A.	1 × 6-pin
GeForce GTX 745	GM107	055	k. A.	k. A.	k. A.	no
GeForce GTX 750	GM107	055	7-9	56-59	76	no
GeForce GTX 750 Ti	GM107	060	4-8	52-60	60-66	no
GeForce GTX 760 192-bit	GK104	130	k. A.	k. A.	k. A.	1 × 6-pin
GeForce GTX 760	GK104	170	10-12	148-169	165-197	2 × 6-pin
GeForce GTX 760 Ti	GK104	170	k. A.	k. A.	k. A.
GeForce GTX 770	GK104	230	8-10	162-195	213-214	1 × 6-pin 1 × 8-pin
GeForce GTX 780	GK110 GK110B	250	10-14	199-222	252-268
GeForce GTX 780 Ti	GK110B	250	10-16	229-260	260-276
GeForce GTX Titan	GK110	250	10-13	198-214	250-263
GeForce GTX Titan Black	GK110B	250	13	243	k. A.
GeForce GTX Titan Z	2 × GK110B	375	k. A.	k. A.	k. A.	2 × 8-pin

Remarks

1. ↑ The date indicated is the date of the public presentation, not the date of availability of the models.
2. ↑ The specified performance values ​​for the computing power via the stream processors, the pixel and texel filling rate, as well as the memory bandwidth are theoretical maximum values ​​(with standard clock rate) that cannot be directly compared 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.
3. ↑ ^{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. 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.
4. ↑ The shader units of the GF108 graphics processor are clocked twice as high in relation to the rest of the chip.
5. ↑ ^{a b c} OEM product. Card is not available in the retail market.
6. ↑ 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.
7. ↑ 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.
8. ↑ The value given under 3D load corresponds to the typical game usage of the card. However, this is different depending on the 3D application. As a rule, a modern 3D application is used to determine the value, which, however, limits the comparability over longer periods of time.
9. ↑ The maximum load is usually determined with demanding benchmark programs, the loads of which are significantly higher than those of "normal" 3D applications.

Web links

Commons : Nvidia GeForce 700 series  - collection of images, videos and audio files

• Product overview on the manufacturer's website

Individual evidence

1. ↑ Review: Nvidia GeForce GTX Titan. Computerbase, February 21, 2013, accessed March 29, 2014 . 
2. ↑ Nvidia GeForce GTX 780 in the test. Computerbase, May 23, 2013, accessed March 29, 2014 . 
3. ↑ Nvidia GeForce GTX 780 Ti versus GTX Titan in the test. Computerbase, November 8, 2013, accessed March 29, 2014 . 
4. ↑ Nvidia GeForce GTX Titan Black with 6 GB in the test. Computerbase, March 7, 2014, accessed March 29, 2014 . 
5. ↑ Nvidia GeForce GTX Titan Z with 8 TFLOPS for $ 3,000. Computerbase, March 25, 2014, accessed March 29, 2014 . 
6. ↑ Nvidia releases the GeForce GTX Titan Z. Computerbase, May 28, 2014, accessed May 28, 2014 . 
7. ↑ Nvidia GeForce GTX 770 in the test. Computerbase, May 30, 2013, accessed March 29, 2014 . 
8. ↑ Nvidia GeForce GTX 760 in the test. Computerbase, June 25, 2013, accessed March 29, 2014 . 
9. ↑ Nvidia GeForce GTX 750 (Ti) "Maxwell" in the test. Computerbase, February 18, 2014, accessed March 29, 2014 . 
10. ↑ http://www.geeks3d.com/20150519/nvidia-r352-86-whql-released-one-new-opengl-extension/ Support for OpenGL 4.5 and OpenCL 1.2 with driver 35x.xx and higher for Maxwell and Kepler
11. ↑ https://www.techpowerup.com/gpudb/1986/geforce-gtx-750
12. ↑ NVIDIA GeForce GT 710. Nvidia Corporation, accessed April 28, 2018 . 
13. ^ NVIDIA GeForce GT 720. Nvidia Corporation, accessed September 19, 2014 . 
14. ↑ ^{a b c} NVIDIA GeForce GT 730. Nvidia Corporation, accessed June 19, 2014 . 
15. ↑ NVIDIA GeForce GT 740. Nvidia Corporation, accessed May 30, 2014 . 
16. ↑ NVIDIA GeForce GTX 745 (OEM). Nvidia Corporation, accessed June 26, 2014 . 
17. ↑ NVIDIA GeForce GTX 750. Nvidia Corporation, accessed February 18, 2014 . 
18. ↑ NVIDIA GeForce GTX 750 Ti. Nvidia Corporation, accessed February 18, 2014 . 
19. ↑ NVIDIA GeForce GTX 760 192-bit (OEM). Nvidia Corporation, accessed October 20, 2013 . 
20. ↑ NVIDIA GeForce GTX 760. Nvidia Corporation, accessed June 25, 2013 . 
21. ↑ NVIDIA GeForce GTX 760 Ti (OEM). Nvidia Corporation, accessed October 20, 2013 . 
22. ↑ NVIDIA GeForce GTX 770. Nvidia Corporation, accessed May 30, 2013 . 
23. ↑ NVIDIA GeForce GTX 780. Nvidia Corporation, accessed May 23, 2013 . 
24. ↑ NVIDIA GeForce GTX 780 Ti. Nvidia Corporation, accessed November 7, 2013 . 
25. ↑ NVIDIA GeForce GTX TITAN. Nvidia Corporation, accessed June 4, 2013 . 
26. ↑ NVIDIA GeForce GTX TITAN Black. Nvidia Corporation, accessed February 18, 2014 . 
27. ↑ NVIDIA GeForce GTX TITAN Z. Nvidia Corporation, accessed May 28, 2014 . 
28. ↑ ^{a b c d e f g h i j k l m n o p q r} ASUS GTX 750 OC 1 GB. TechPowerUp, February 19, 2014, accessed April 24, 2014 . 
29. ↑ ^{a b} Nvidia lance les GeForce GTX 750 Ti & 750 Maxwell. Hardware.fr, February 18, 2014, accessed April 24, 2014 (French). 
30. ↑ ^{a b c d e f g h i j k l m n} NVIDIA GeForce GTX 750 Ti in the test (page 44). HardTecs4U, February 18, 2014, accessed April 24, 2014 . 
31. ↑ ^{a b} GeForce GTX 760 in the test: New star in the price / performance sky? PC Games Hardware, June 29, 2013, accessed April 24, 2014 . 
32. ↑ ^{a b c d e} Nvidia GeForce GTX 780 Ti in the test: The "Titan Ultra" has landed - for 649 euros. PC Games Hardware, November 8, 2013, accessed April 24, 2014 . 
33. ^ ^{A b} Hochadel: NVIDIA GeForce GTX Titan Black in the test (page 15). HardTecs4U, June 10, 2014, accessed September 19, 2014 .