Around 3 weeks after the GeForce GT 740, NVIDIA releases today another level entry graphics card: the GeForce GT 730. The GT 730 is based on the the GK208 GPU (Kepler architecture) for the 64-bit versions and on a GF108 (Fermi architecture) for the 128-bit version. This card is available in three versions, with important differences between versions which make the choice of this almost unworthy card difficult.
GeForce GT 730 1GB DDR3 128-bit
- GPU: GF108 @ 700MHz
- CUDA cores: 96 (2 SMs, 48 cores/SM)
- Memory: 1024MB DDR3 128-bit, 1800MHz
- Max digital resolution: 2560 x 1600
- TDP: 49W
- 3D APIs: OpenGL 4.4, Direct3D 11
GeForce GT 730 2GB DDR3 64-bit
- GPU: GK208 @ 902MHz
- CUDA cores: 384 (2 SMX, 192 cores/SMX)
- Memory: 2048MB DDR3 64-bit, 1800MHz
- Max digital resolution: 2560 x 1600
- TDP: 23W
- 3D APIs: OpenGL 4.4, Direct3D 11
GeForce GT 730 1GB GDDR5 64-bit
- GPU: GK208 @ 902MHz
- CUDA cores: 384 (2 SMX, 192 cores/SMX)
- Memory: 1024MB GDDR5 64-bit, 5000MHz effective (1250MHZ real speed)
- Max digital resolution: 2560 x 1600
- TDP: 38W
- 3D APIs: OpenGL 4.4, Direct3D 11
GT 730 64bit GDDR5 & DDR3 is actually GK208, JeGX.
Their specs are exactly the same as the GT 640 GDDR5 and the GT 630 (Kepler) except lower core clock on the GDDR5 card compared to GT 640 GDDR5.
http://www.geforce.com/hardware/desktop-gpus/geforce-gt640/specifications
http://www.geforce.com/hardware/desktop-gpus/geforce-gt-630/specifications
The GT 730 with 96 CUDA cores is a Fermi GF108 GT 430 rebrand.
Thanks, updated!
JeGX, GF108 has 48 cores per SM, so only 2 SM.
GF100/110 has 32 cores per SM, while GF104/114, GF106/116 and GF108 has 48 cores per SM.
Also “Memory: 2048MB DDR3 64-bit, 5000MHz effective (1250MHZ real speed)” should be GDDR5 not DDR3.
bad cpy/past for the GDDR5 mem!
Low profile version?
Is a low power 710 version coming?
the GeForce GT 730 1GB GDDR5 64-bit would be nice if the bus width was bigger..64bit is crippling
Does anybody know what’s happening with a Maxwell project? NVIDIA released just one low-end GPU (GM107) in February, and after that … nothing. 🙁
@Aleksander Probably the usual fab problems which they blame anybody but themselves for