RTX 4090
RTX 4090 Review & Specifications
The RTX 4090 is based on the Ada Lovelace architecture and manufactured using the 5 nm nm process node. It was launched approximately 3 years, 9 months ago, originally priced at $1599. It features 16384 cores, including 128 compute units (SMs), 512 texture mapping units (TMUs), and 176 render output units (ROPs). For ray tracing and AI-accelerated tasks, it also includes 128 dedicated RT cores and 512 Tensor cores, enabling advanced rendering and upscaling features where supported. It has a base clock of 2235 MHz and a boost clock reaching up to 2520 MHz. The RTX 4090 comes equipped with 24 GB of GDDR6X memory across a 384-bit bus. Memory runs at 1313 MHz and delivers a total memory bandwidth of 1 GB/s.
This GPU has a very high thermal design power (TDP) of 450 W. A PC case with excellent airflow and a high-quality, high-wattage power supply are essential for stable operation.
RTX 4090 Performance
RTX 4090 is an ultra high-end graphics card, achieving a relative gaming performance score of 88.46%. It delivers exceptional 4K performance and is ideal for premium builds with a high-end CPU. Expect 300+ FPS in competitive games or over 100 FPS in AAA titles at 4K with ultra settings.
In terms of memory capacity, the RTX 4090 features an exceptionally high VRAM size of 24 GB. This allows users to push the limits of current-generation software and games. You can comfortably run modern AAA titles at ultra settings with full ray tracing enabled at 4K or 5K resolutions, even while using texture packs, high-resolution assets, and real-time lighting. The GPU also excels in professional workloads such as 3D modeling, video editing at 8K, high-resolution texture baking, and multi-layer compositing in tools like Blender, DaVinci Resolve, or Adobe After Effects. In the domain of AI and machine learning, this level of VRAM supports complex neural networks (e.g., large LLMs, diffusion models like Stable Diffusion XL) with high batch sizes, and seamless multitasking without swapping to system memory. This GPU is also ideal for creators who run multiple high-memory applications simultaneously.
RTX 4090 has a gaming performance-per-price ratio of 24.1%, suggesting lower value in this category. However, its high raw performance makes it a viable choice for users who prioritize gaming power over budget constraints.
General Information
| GPU Name | RTX 4090 |
|---|---|
| GPU Code Name | AD102 |
| Release Price | $1599 |
| Selected Model | GeForce RTX 4090 Gaming OC 24GB Graphics Card - 24GB GDDR6X, PCI-E 4.0, Core 2535Mhz, RGB Fusion, Anti-sag Bracket, Metal Back Plate, DP 1.4, HDMI 2.1a, NVIDIA DLSS 3, GV-N4090GAMING OC-24GD |
| Current Price Amazon | $3400 |
| Amazon Price Updated | 1 week ago |
| Launch Date | 3 years, 9 months ago |
| GPU Available | True |
| Architecture | Ada Lovelace |
|---|---|
| Foundry | TSMC |
| Process | 5 nm |
| Transistors | 76300 million |
| Die Size | 609 mm² |
| Bus Interface | PCIe 4.0 x16 |
| TDP | 450 W |
| Cores | 16384 |
|---|---|
| CUs (SMs) | 128 |
| TMUs | 512 |
| ROPs | 176 |
| RT Cores | 128 |
| Tensor Cores | 512 |
| L1 Cache | 128 KB (per SM) |
| L2 Cache | 72 MB |
| Base Clock | 2235 MHz |
|---|---|
| Boost Clock | 2520 MHz |
| Memory Size - Type | 24 GB GDDR6X |
|---|---|
| Memory Bus Width | 384 bit |
| Memory Clock | 1313 MHz |
| Memory Bandwidth | 1 GB/s |
| DirectX | 12 Ultimate (12_2) |
|---|---|
| OpenGL | 4.6 |
| OpenCL | 3.0 |
| Vulkan | 1.4 |
| Shader Model | 6.8 |
Theoretical Performance
| Pixel Rate | 443.50 GPixel/s |
|---|---|
| Texture Rate | 1290.20 GTexel/s |
| FP16 | 82.58 TFLOPS |
| FP32 | 82.58 TFLOPS |
| FP64 | 1290.22 GFLOPS |
Gaming Performance Chart
RTX 4090 Compared to Similar Performance GPUs
GPU Price / Performance Table
| GPU (Desc) | Price (Desc) | Gaming_1080p (Desc) | Gaming_1440p (Desc) | Gaming_2160p (Desc) | Game_1440p/Price (Desc) |
|---|---|---|---|---|---|
|
$12380
|
0.0%
|
103.1%
|
106.1%
|
110.1%
|
8.5%
|
|
|
$4900
|
0.0%
|
100.0%
|
100.0%
|
100.0%
|
20.2%
|
|
|
$3400
|
0.0%
|
88.5%
|
82.9%
|
75.2%
|
24.1%
|
|
|
$1450
|
0.0%
|
76.9%
|
72.0%
|
65.1%
|
49.1%
|
|
|
$1460
|
↓
-4.8%
|
74.6%
|
67.1%
|
58.7%
|
45.5%
|
|
|
$1460
|
↓
-4.8%
|
73.9%
|
66.8%
|
58.3%
|
45.3%
|
|
|
$1400
|
0.0%
|
72.9%
|
66.1%
|
59.8%
|
46.7%
|
|
|
$1039
|
↓
-3.9%
|
70.0%
|
64.0%
|
56.9%
|
61.0%
|
|
|
$740
|
↓
-4.1%
|
69.2%
|
62.3%
|
54.1%
|
83.4%
|
GPU Specification Table
| GPU (Desc) | Architecture | Shader/TMU/ROP/RT | Base/Boost Clock | Memory | Mem Clock/Bandwidth | TDP |
|---|---|---|---|---|---|---|
| RTX PRO 6000 Blackwell | GB202 / 5 nm nm / Blackwell 2.0 | 24064 / 752 / 192 / 188 | 1590 MHz / 2617 MHz | 96 GB GDDR7 / 512 bit | 1750 MHz / 1 GB/s | 600 W |
| RTX 5090 | GB202 / 5 nm nm / Blackwell 2.0 | 21760 / 680 / 176 / 170 | 2017 MHz / 2407 MHz | 32 GB GDDR7 / 512 bit | 1750 MHz / 1 GB/s | 575 W |
| RTX 4090 | AD102 / 5 nm nm / Ada Lovelace | 16384 / 512 / 176 / 128 | 2235 MHz / 2520 MHz | 24 GB GDDR6X / 384 bit | 1313 MHz / 1 GB/s | 450 W |
| RTX 5080 | GB203 / 5 nm nm / Blackwell 2.0 | 10752 / 336 / 112 / 84 | 2295 MHz / 2617 MHz | 16 GB GDDR7 / 256 bit | 1875 MHz / 960 GB/s | 360 W |
| RTX 4080 SUPER | AD103 / 5 nm nm / Ada Lovelace | 10240 / 320 / 112 / 80 | 2295 MHz / 2550 MHz | 16 GB GDDR6X / 256 bit | 1438 MHz / 736 GB/s | 320 W |
| RTX 4080 | AD103 / 5 nm nm / Ada Lovelace | 9728 / 304 / 112 / 76 | 2205 MHz / 2505 MHz | 16 GB GDDR6X / 256 bit | 1400 MHz / 716 GB/s | 320 W |
| RX 7900 XTX | Navi 31 / 5 nm / RDNA 3.0 | 6144 / 384 / 192 / 96 | 1929 MHz / 2498 MHz | 24 GB GDDR6 / 384 bit | 2500 MHz / 960 GB/s | 355 W |
| RTX 5070 Ti | GB203 / 5 nm nm / Blackwell 2.0 | 8960 / 280 / 96 / 70 | 2295 MHz / 2452 MHz | 16 GB GDDR7 / 256 bit | 1750 MHz / 896 GB/s | 300 W |
| RX 9070 XT | Navi 48 / 5 nm / RDNA 4.0 | 4096 / 256 / 128 / 64 | 1660 MHz / 2970 MHz | 16 GB GDDR6 / 256 bit | 2518 MHz / 644 GB/s | 304 W |
Add Comment on RTX 4090