AMD Radeon RX 6600 8GB Specs, Benchmarks & Pricing
The AMD Radeon RX 6600 was released in October 2021 as an affordable mainstream gaming GPU based on the RDNA 2 architecture (Navi 23 XL die). It features 28 compute units with 1,792 stream processors and 8GB of GDDR6 memory with 224 GB/s bandwidth on a 128-bit bus. The card includes 28 ray tracing acceleration cores and 32MB of Infinity Cache. Built on TSMC's 7nm process with a 237 mm² die containing 11.06 billion transistors, it delivers 8.93 TFLOPS of FP32 performance at a boost clock of 2,491 MHz. With a TDP of only 132W, the RX 6600 was designed as AMD's most efficient 1080p gaming card, offering excellent performance per watt at a lower price point than the RX 6600 XT. It competes with NVIDIA's RTX 3060 and offers solid 1080p gaming with DirectX 12 Ultimate support including hardware-accelerated ray tracing. It uses a PCIe 4.0 x8 interface.
- Release Date: October 13, 2021
- MSRP: $329 USD
- GPU Architecture: RDNA 2
- Hardware-Accelerated GEMM Operations:FP64 FP32 TF32 BF16 FP16 FP8 FP4 INT8 INT4
- CUDA Compute Capability : n/a
Considerations
- Lower FP32 compute performance compared to other GPUs
- Lower FP16 compute performance compared to other GPUs
- Below average 3DMark Wild Life Extreme (4K) performance
- Below average 3DMark Wild Life Extreme (1440p) performance
Specifications for AMD Radeon RX 6600
| Specification | Performance Ranking |
|---|---|
| FP32 TFLOPs | |
| FP16 TFLOPs | |
| Tensor Core Count | Data not available |
| Memory Capacity (GB) | |
| Memory Bandwidth (GB/s) | |
| Int8 TOPs | Data not available |
| FP8 TFLOPs | Data not available |
| FP4 TFLOPs | Data not available |
Gaming Benchmarks for AMD Radeon RX 6600
Rankings based on 805,683 real-world benchmark results from OpenBenchmarking.org
| Benchmark | Performance Ranking |
|---|---|
| 3DMark Wild Life Extreme (4K) | |
| 3DMark Wild Life Extreme (1440p) | |
| 3DMark Wild Life Extreme (1080p) | |
| FurMark (GL) (4K) | |
| FurMark (GL) (1440p) | |
| FurMark (GL) (1080p) | |
| FurMark (VK) (4K) | |
| FurMark (VK) (1440p) | |
| FurMark (VK) (1080p) | |
| Quake II RTX (4K) |
Real-time AMD Radeon RX 6600 GPU Prices
Compare Price/Performance to other GPUs
Compare AMD Radeon RX 6600 to Another GPU
Price History
AMD Radeon RX 6600 Price History
This chart tracks AMD Radeon RX 6600 prices over the past 6 months based on eBay listings. The yellow line shows the average of the three lowest-priced listings each day—a useful indicator for finding deals.
Trend: Over this period, the lowest average price has been relatively stable. The current lowest average is $167, compared to a typical $181 over the period. The lowest average price is calculated as the average of the three lowest-priced listings each day.
References
- https://www.techpowerup.com/gpu-specs/radeon-rx-6600.c3696
- https://www.amd.com/en/products/graphics/desktops/radeon/6000-series/amd-radeon-rx-6600.html
- https://www.amd.com/en/technologies/rdna.html
- https://www.tomshardware.com/reviews/amd-radeon-rx-6600-review-xfx
- https://en.wikipedia.org/wiki/RDNA_2
- https://www.anandtech.com/show/17007/amd-launches-radeon-rx-6600-more-mainstream-gaming-for-329
Notes
- FP32 TFLOPS calculated as 1,792 stream processors × 2,491 MHz boost clock × 2 ops/cycle / 1,000,000 = 8.93 TFLOPS
- FP16 performance is 17.86 TFLOPS calculated at 2:1 ratio to FP32, which is standard for RDNA 2 architecture
- Memory bandwidth of 224 GB/s calculated as 128-bit bus × 14 Gbps effective / 8
- RDNA 2 architecture does not include dedicated AI accelerators or tensor cores (added in RDNA 3), so tensorCoreCount is null
- supportedHardwareOperations is an empty array: RDNA 2 has no matrix/tensor-core (GEMM-accelerated) hardware path of any kind. Per AMD's own RDNA architecture technology page, the generational feature-comparison table's "AI Accelerators" row is blank for both the "RDNA" and "RDNA 2" columns and first shows "1st Gen" at "RDNA 3" -- confirmed independently by Wikipedia's RDNA 2 article, which makes no mention of AI accelerators or matrix-multiply hardware. RDNA 2's ray accelerator is a distinct fixed-function ray-intersection unit, unrelated to matrix/GEMM support. Sources: https://www.amd.com/en/technologies/rdna.html, https://en.wikipedia.org/wiki/RDNA_2
- fp8TFLOPS and fp4TFLOPS are left null and intentionally NOT listed in unverifiedSpecs: this is a confirmed hardware negative rather than a research gap. RDNA 2 predates AMD's AI Accelerators/WMMA matrix engines entirely (introduced in RDNA 3), so no FP8 or FP4 matrix throughput exists on this architecture to report.
- RDNA 2 CUs execute INT8 and INT4 operations via standard ALU/dot-product instructions, not a dedicated matrix or tensor core -- RDNA 2 predates AMD's AI Accelerators (WMMA), which were introduced in RDNA 3. AMD does not publish INT8 TOPS for consumer RDNA 2 GPUs, so int8TOPS is null. Sources: https://en.wikipedia.org/wiki/RDNA_2, https://www.amd.com/en/technologies/rdna.html
- The 32MB Infinity Cache significantly improves effective memory bandwidth beyond the raw 224 GB/s specification
- Built on TSMC 7nm process with 237 mm² die size containing 11.06 billion transistors
- Uses the Navi 23 XL die with 4 compute units disabled compared to the RX 6600 XT
- Uses PCIe 4.0 x8 interface instead of x16
- MSRP of $329 USD verified from multiple sources including Tom's Hardware, AnandTech, and GSMArena at launch on October 13, 2021