Kunlunxin R200-8F vs Kunlunxin R200
The R200-8F and the R200 deliver near-identical FP16 throughput (128 vs 128 TFLOPS dense). The R200-8F also carries 16 GB more memory (32 GB vs 16 GB).
Kunlunxin R200-8F
Full specs →Kunlunxin R200
Full specs →Specifications
Performance (TFLOPS)
FLOPS by Precision
What actually differs
Dense throughput for the R200-8F against the R200: FP32 32 vs 32 TFLOPS, FP16 128 vs 128 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.
Memory is 32 GB against 16 GB, fed at 512 GB/s versus 512 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.
Power budgets are 160 W for the R200-8F and 150 W for the R200. At FP32 that works out to 0.20 against 0.21 TFLOPS per watt.
For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.
Frequently asked questions
Is the R200-8F faster than the R200?
They are close on paper: both deliver about 128 TFLOPS of dense FP16 throughput. Memory, bandwidth, and power are the deciding differences.
Which has more memory, the R200-8F or the R200?
The R200-8F carries 32 GB of memory versus 16 GB for the R200. Memory bandwidth is 512 GB/s for the R200-8F and 512 GB/s for the R200.
How much power do the R200-8F and the R200 draw?
The R200-8F is rated at 160 W TDP and the R200 at 150 W. On FP32 throughput per watt, the R200 is the more efficient part.
Get Comparison Updates
New GPUs added weekly. Be the first to see how they compare.