I thought it be good to put the specs of my local GPU so that in future, any tests that I do I can use this as a reference.
Make: NVIDIA GTX 1050
This GPU is built using NVIDIA’s Pascal architecture, a successor of Maxwell which was introduced in 2016.
The GPU’s base frequency is 1354 MHz; the GPU Boost rating jumps to 1455 MHz.
Like the 1050 Ti, GeForce GTX 1050 exposes 32 ROPs and the same 1 MB of L2. Instead of 4 GB on its aggregate 128-bit bus, though, Nvidia deploys 2 GB at the same 7 GT/s.
Theoretical maximum bandwidth remains the same at 112 GB/s.
- GPU Architecture: Pascal
- CUDA Cores: 640
- Standard Memory: 2GB GDDR5
- Memory Speed: 7GB/s
- Memory Interface Width: 128 bit
- Memory Bandwidth (GB/s): 112
- Base Clock: 1354 Mhz
- Boost Clock: 1455 Mhz
- NVDIA G-Synch
Display & Resolution
- Max Resolution 7680*4320@60Hz
- Ports: DP1.4, HDMI 2.0b, Duallink-DVI
- Multimonitor: Yes
- HDCP: 2.2
Power & Thermal
The GPU has a tiny TDP (Thermal Design Power) of just 75W)
- Max GPU Temperarture: 95 Degrees Centigrade
- Card Power: 75W
- Recomended Power Supply: 300W+
What are 640 CUDA Cores?
CUDA is a parallel computing platform and application programming interface (API) model created by Nvidia, which allows software developers and software engineers to use a CUDA-enabled graphics processing unit (GPU) for general purpose processing. GPU Cores are parrarel processors responsible for processing all the data that is fed in and out of the GPU.
CUDA cores on the other hand are similar to CPU cores BUT cannot fetch instructions, cannot decode instructions, cannot access memory. CAN process large datasets. CPU’s have several cores but GPU have thousands or cores.
An NVIDIA GPU that has more CUDA cores does not mean that it more powerful than one with less CUDA cores. This is only true if they are from a similar family.
Here is where thing become interesting.
The GPU Architecture as seen above has 16 streaming multiprocessors, aka SMwhich contains 32 CUDA cores each. Note that every CUDA is an execute unit for integer and fl>
So in simpler words, 16SM x 32 = 512 CUDA cores.
512 is the minimum a GPU can have, mine is at 640 so its well above 512.