AMD's reference Radeon R9 290 and 290X cooler earned our disdain once we figured out that it was causing the otherwise-impressive Hawaii GPU to operate at performance levels below the company's specification. Getting the graphics processor to deliver full performance required faster fan speeds, more noise, and ultimately higher power.
Starting back in December, partner boards with third-party coolers started showing up. No longer was it necessary to make your own modifications, like we did in Fixing The Radeon R9 290 With Arctic's Accelero Xtreme III, to maximize the performance of those once-$400 and -$550 cards. Speaking of, let's take a look at how the cards in today's update are priced now:
| Radeon R9 290 (Newegg) | |
|---|---|
| Sapphire Tri-X OC R9 290 | $650 |
| Gigabyte GV-R929OC-4GD R9 290 Windforce OC | Out of stock |
| Radeon R9 290X (Newegg) | |
|---|---|
| Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC | $730 |
| Sapphire Tri-X OC R9 290X | Out of stock |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC | $700 |
| HIS R9 290X IceQ X² Turbo | Not Available |
| MSI R9 290X Gaming 4G | $700 |
Our colleagues in Germany began working with the uniquely-tuned offerings, hoping to find that AMD's partners were extracting just as much performance as we were able to, without the hassle of adding your own aftermarket cooler. Do they, though? That's the question we want to answer today.
In Germany, this story has been taking shape one card at a time. But now that we have it translated into English, we have a more complete picture of the partner card market, including two Radeon R9 290 cards, five R9 290X cards, the reference board with our Arctic Accelero Extreme III modification, and another home-grown configuration with NZXT's Kraken G10 and X40 bracket.
All of the Radeon R9 290X cards are set to their Quiet Mode firmware setting, since the board partners do a suitable job of cooling Hawaii without sacrificing clock rates.
Radeon R9 290 Model Overview
| Sapphire Tri-X OC R9 290 | ![]() |
|---|---|
| Radeon R9 290 + Arctic Accelero Extreme III | ![]() |
| Radeon R9 290 + NZXT Kraken G10 + X40 | |
| Gigabyte R9 290 Windforce OC (Update) | ![]() |
Radeon R9 290X Model Overview
| Asus R9 290X DirectCU II OC | ![]() |
|---|---|
| Sapphire Tri-X OC R9 290X | ![]() |
| Gigabyte R9 290X Windforce OC | ![]() |
| HIS R9 290X IceQ X² Turbo (Update) | ![]() |
| MSI R9 290X Gaming 4G (Update) | ![]() |
All of these cards are based on the same Hawaii GPU manufactured at 28 nm with either 2560 (in the case of the Radeon R9 290) or 2816 (for the Radeon R9 290X) shaders and a 512-bit memory bus. The differences between these cards, aside from their coolers, are limited to the core and memory clock frequencies.
| Radeon R9 290 | GPU Clock MHz (Boost) | Memory Clock MHz | Memory Bandwidth GB/s | Pixel Fillrate GPixel/s | Texture Fillrate GTexel/s |
|---|---|---|---|---|---|
| Sapphire Tri-X OC R9 290 | 1000 | 1300 | 332.8 | 64.0 | 160.0 |
| Gigabyte GV-R929OC-4GD R9 290 Windforce OC | 1040 | 1250 | 320.0 | 66.6 | 166.4 |
| Radeon R9 290 Reference + Arctic Accelero Extreme III | 1100 | 1250 | 320.0 | 70.4 | 176.0 |
| Radeon R9 290 Reference + NZXT Kraken G10 + X40 | 1100 | 1250 | 320.0 | 70.4 | 176.0 |
| Radeon R9 290X | GPU Clock MHz (Boost) | Memory Clock MHz | Memory Bandwidth GB/s | Pixel Fillrate GPixel/s | Texture Fillrate GTexel/s |
|---|---|---|---|---|---|
| Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC | 1050 | 1350 | 345.6 | 67.2 | 184.8 |
| Sapphire Tri-X OC R9 290X | 1040 | 1300 | 332.8 | 66.6 | 183.0 |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC | 1040 | 1250 | 320.0 | 66.6 | 183.0 |
| HIS R9 290X IceQ X² Turbo | 1060 | 1350 | 345.6 | 67.8 | 186.6 |
| MSI R9 290X Gaming 4G | 1040 | 1250 | 320 | 66.6 | 183 |
Of course, there’s more to the overall performance story than just the maximum GPU clock rate. As we've already seen, calling your GPU a 1000 MHz part is pointless if it cannot maintain that frequency. This is precisely the problem AMD's reference design suffers from. Once its target temperature is hit, the clock rate starts scaling back and performance follows suit.
We warmed up our various contenders and ran them in a loop while we recorded their frequencies.
R9 290 GPU Clock Frequency




R9 290X GPU Clock Frequency





The dimensions reported here don't necessarily match what you've heard from each manufacturer's official technical specifications. Rather, we measure them by hand to assure they're correct. The image and chart below should help illustrate what each measurement actually means. Auxiliary PCI Express power connectors are not included; they have to be added depending on the power plug and cable design.

| Radeon R9 290 | Length (L) | Height (H) | Depth (D1) | Depth (D2) |
|---|---|---|---|---|
| Sapphire Tri-X OC R9 290 | 305 mm | 114 mm | 38 mm | 4 mm |
| Gigabyte GV-R929OC-4GD R9 290 Windforce OC | 282 mm | 123 mm | 38 mm | 4 mm |
| Radeon R9 290 Reference + Arctic Accelero Extreme III | 320 mm | 120 mm | 60 mm | 4 mm |
| Radeon R9 290X | Length (L) | Height (H) | Depth (D1) | Depth (D2) |
|---|---|---|---|---|
| Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC | 288 mm | 142 mm | 38 mm | 4 mm |
| Sapphire Tri-X OC R9 290X | 305 mm | 114 mm | 38 mm | 4 mm |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC | 282 mm | 123 mm | 38 mm | 4 mm |
| HIS R9 290X IceQ X² Turbo | 297 mm | 135 mm | 36 mm | 4 mm |
| MSI R9 290X Gaming 4G | 279 mm | 120 mm | 38 mm | 6 mm |
Graphics Card Weight
The weight of a card might be interesting if you're trying to figure out if any additional support is needed, or to calculate the amount of stress your motherboard might be under in a CrossFire-based setup.
| Radeon R9 290 | |
|---|---|
| Sapphire Tri-X OC R9 290 | 1022 g |
| Gigabyte GV-R929OC-4GD R9 290 Windforce OC | 1040 g |
| Radeon R9 290 Reference + Arctic Accelero Extreme III | 978 g |
| Radeon R9 290X | |
|---|---|
| Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC | 1135 g |
| Sapphire Tri-X OC R9 290X | 1022 g |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC | 1053 g |
| HIS R9 290X IceQ X² Turbo | 976 g |
| MSI R9 290X Gaming 4G | 1038 g |
Benchmark System And Procedure
We collaborated with HAMEG (Rohde & Schwarz) to upgrade our power consumption measurement system.

We record all channels and the corresponding oscilloscope value/curves for our measurements. The very precise and, more important, fast current clamps yield 100 mV/A, making it easy to calculate the power based on the voltage. We also record the supply voltage to multiply its value with the recorded amperage. Depending on the resolution we choose, this procedure yields a very detailed power consumption history. We generally set this to 1 ms, allowing us to capture all fluctuations attributable to AMD’s PowerTune or Nvidia’s GPU Boost technology.
| Measurement Procedure | Non-Contact Direct Current Measurement at the PCIe Slot Non-Contact Direct Current Measurement at the External PCIe Power Supply Direct Voltage Measurement 3.3 V / 12 V |
|---|---|
| Measurement Apparatus | Oscilloscope: Power Clamp: Voltage Divider Probe: Digital Multimeter: |
| Bench Table | Microcool Banchetto 101 |
| Test Hardware | AMD FX-8350 (Piledriver), Overclocked to 4.5 GHz Corsair H100i Compact Water Cooling Solution 16 GB (2 x 8) Corsair Vengeance DDR3-1866 Asus 990FX Sabertooth + Modified PCIe Adapter with Current Loops |
| Power Supply | Corsair AX860i with Modified Plugs (Pickup) |
Power Consumption While Running A Gaming Loop
"How much power does a graphics card draw during gaming?" and "How much heat does it generate under load?" are the most commonly asked questions once we wrap up our analysis of 3D performance. Our testing is made as real-world as possible by measuring cards that have already been warmed up.

This high-resolution measurement shows why power supplies can be overwhelmed unless they have ample output headroom. Even if a PSU's specs suggest it should be able to handle a given card, some very brief (often less than 10 ms), but very high peaks can cause a power supply's protection circuitry to engage.
Power Consumption: Radeon R9 290




Power Consumption: Radeon R9 290X





Benchmark System and Software
When I wrote Seven Radeon R9 280X Graphics Cards, Rounded-Up, I made the assumption that most of our readers would know that gaming performance would correlate to clock rates, and focused most of my efforts on power, heat, and noise. We heard the feedback from that piece, though, and incorporated benchmark results into today's story.
I ran four carefully-selected titles at their highest quality settings, and then normalized and added the results. This yielded a performance index with AMD's Radeon R9 290 reference card serving as the 100-percent baseline.

| CPU And Cooler | Intel Core i7-3770K (Ivy Bridge) at 4.5 GHz Corsair H100i Compact Water Cooler (Gelid GC Extreme) |
|---|---|
| Motherboard | Gigabyte G1. Sniper 3 (Z77 Express) |
| Memory | 32 GB (4 x 8 GB) Corsair Dominator Platinum DDR3-2133 |
| SSD | 2 x Corsair Neutron 480 GB |
| Power Supply | Corsair AX1200i |
| Operating System | Windows 7 x64 Ultimate SP1 |
| Drivers | AMD Catalyst 13.12 GeForce 331.82 |
| Benchmarks | Metro: Last Light BioShock Infinite Battlefield 4 (Single-Player) Crysis 3 (DirectX 11) |
We warmed up each graphics card until it reached its peak temperature, keeping us as close as possible to real-world performance and to prevent unfair advantages for either AMD or Nvidia due to artificially high boost speeds. The cards were tested on an open bench table (blue bars) and in a closed mid-tower (red bars).

We're using Metro: Last Light for our thermal benchmarks. This should be a good representation of a high-end graphics workload. As a bonus, it's incredibly repeatable and easy to loop over and over through the built-in tool.
After publishing Does Radeon R9 290X Behave Any Differently In A Closed Case?, we decided to test these cards on an open bench table and in a closed case, since we found that dropping them into the right enclosure helped lower temperatures thanks to optimized airflow.
| Model | Idle | Gaming, Open-Air Bench | VRM | Gaming, Closed Case |
|---|---|---|---|---|
| Sapphire Tri-X OC R9 290 | 32 °C | 72 °C | 78 °C | 60-71 °C |
| Gigabyte GV-R929OC-4GD R9 290X Windforce OC | 34 °C | 72 °C | 77 °C | 73-74 °C |
| Radeon R9 290 Reference + Arctic Accelero Extreme III | 30 °C | 68 °C | 90 °C | 72 °C |
| Radeon R9 290 Reference + NZXT Kraken G10 + X40 | 28 °C | 49 °C | 66 °C | 49 °C |
| Model | Idle | Gaming, Open-Air Bench | VRM | Gaming, Closed Case |
|---|---|---|---|---|
| Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC | 34 °C | 76 °C | 92 °C | 84-85 °C |
| Sapphire Tri-X OC R9 290X | 35 °C | 73 °C | 85 °C | 70-72 °C |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC Press Sample | 34 °C | 84 °C | 86 °C | 83 °C |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC Mass Production | 34 °C | 83 °C | 87 °C | 81 °C |
| HIS R9 290X IceQ X² Turbo | 35 °C | 78 °C | 70 °C | 81-82 °C |
| MSI R9 290X Gaming 4G | 34 °C | 76 °C | 73 °C | 75-76 °C |
Temperature Graphs
These graphs slow how the temperatures change over time, influenced by the heat sink, the fan, and the firmware's fan profile.
Asus' R9 290X DirectCU II OC has a bit of a problem; its cooler was actually designed for the larger GK110 GPU on its GeForce GTX 780 Ti board. You can see this in the image below, where two of the heat pipes don't tough Hawaii at all, and two others make partial contact.
Sapphire solves this with a group of three heat pipes that better fit AMD's GPU. The result is made clear in our testing.
Radeon R9 290 Temperatures




Radeon R9 290X Temperatures





Noise
We first measure each graphics card's noise level in different workloads using the same studio microphone and calibration seen in our audio reviews. The microphone is positioned perpendicular to the middle of the graphics card at a distance of 50 cm.
| Radeon R9 290 | Idle | Gaming, Open-Air Bench | Gaming, Closed Case |
|---|---|---|---|
| Sapphire Tri-X OC R9 290 | 31.4 dB(A) | 39.4 dB(A) | 40.5 dB(A) |
| Gigabyte GV-R929OC-4GD R9 290 Windforce OC | 30.9 dB(A) | 39.5 dB(A) | 42.7 dB(A) |
| Radeon R9 290 Reference + Arctic Accelero Extreme III | 37.8 dB(A) | 37.8 dB(A) | 37.8 dB(A) |
| Radeon R9 290 Reference + NZXT Kraken G10 + X40 | 34.2 dB(A) | 37.2 dB(A) | --- |
| Radeon R9 290X | Idle | Gaming, Open-Air Bench | Gaming, Closed Case |
|---|---|---|---|
| Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC | 32.5 dB(A) | 42.3 dB(A) | 44.3 dB(A) |
| Sapphire Tri-X OC R9 290X | 32.1 dB(A) | 40.9 dB(A) | 42.8 dB(A) |
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC Press Sample | 30.9 dB(A) | 41.5 dB(A) | 43.6 dB(A) |
| Gigabyte GV-R929XOC-4GD-4GD R9 290X Windforce OC Mass Production | 30.9 dB(A) | 39.6 dB(A) | 43.2 dB(A) |
| HIS R9 290X IceQ X² Turbo | 31.2 dB(A) | 46.2 dB(A) | 48.8 dB(A) |
| MSI R9 290X Gaming 4G | 30.9 dB(A) | 41.2 dB(A) | 43.9 dB(A) |
Fan Speed
The number, diameter, and rotational speed of the fans are mainly what determine the cooler's noise level. We present the RPM curves for each graphics card separately; we tried, and putting them into one graph is too cluttered.
Both of the models that we modded up (using Arctic's Accelero and NZXT's Kraken) were not measured. Their numbers are not comparable.
Radeon R9 290 Fan Speed


Radeon R9 290X Fan Speed





The combination of how hot each board gets, how well its heat sink is designed, and how fast its fans spin gives each graphics card a unique acoustic signature. The audio track is taken directly from the measurement microphone, which records as we run a long loop of Metro: Last Light.
| Radeon R9 290 |
|---|
Sapphire Tri-X OC R9 290
Gigabyte GV-R929OC-4GD R9 290 Windforce OC
Radeon R9 290 Reference + Arctic Accelero Extreme III
| Radeon R9 290X |
|---|
Asus R9290X-DC2OC-4GD5 R9 290X DirectCU II OC
Gigabyte GV-R929XOC-4GD R9 290X Windforce OC
HIS R9 290X IceQ X² Turbo
MSI R9 290X Gaming 4G
Sapphire Tri-X OC R9 290X
| Radeon R9 290 Reference + Arctic Accelero Extreme III | |
|---|---|
| Auxiliary Power Connectors | 1 x eight-pin + 1 x six-pin |
| Connectors | 2 x DVI-D (Dual-Link, No analog connector) 1 x HDMI 1 x DisplayPort |
| Form Factor | Two Slots |
| Pros | + Relatively light for its size + Both quiet and cool + Fast |
| Cons | - Very long - More work building - Loss of warranty |

Front and Back
Sides
Connectors
Front and Back
Sides
Connectors
| Gigabyte GV-R929XOC-4GD R9 290X Windforce OC | |
|---|---|
| Auxiliary Power Connectors | 1 x eight-pin + 1 x six-pin |
| Connectors | 2 x DVI-D (Dual-Link, No analog connector) 1 x HDMI 1 x DisplayPort |
| Form Factor | Two Slots |
| Pros | + Flat + Relatively short and light |
| Cons | - Runs somewhat warm in AMD's Quiet Mode |
Front and Back
Sides
Connectors











































































