Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsA hardware-modified Radeon RX 7900 XTX reached 3,467 MHz peak and 3,333 MHz average in 3DMark Time Spy, producing a benchmark result close to an RTX 4090 in a selected graphics test. That performance came from bypassing the card’s normal power reporting, drawing as much as 696 W in another test, and using 10°C coolant—not from a simple software setting or an efficiency improvement.
What did the modified RX 7900 XTX achieve?
In a 2023 enthusiast experiment, Reddit user jedi95 connected an Elmor Labs EVC2SE to the graphics card’s I2C bus and adjusted the power value reported by its voltage-regulator controller. With that control ceiling bypassed, the Asus TUF RX 7900 XTX reached clocks far above its advertised boost figure and posted the following 3DMark results:
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
ASRock Radeon RX 7900 XTX Phantom Gaming 24GB OC Graphics Card, 2615 MHz Boost Clock, 24GB GDDR6,... | $1,129.99 | Buy on Amazon |
| Benchmark or measurement | Reported result | Context |
|---|---|---|
| GPU clock | 3,467 MHz peak; 3,333 MHz average | jedi95’s 2023 Time Spy run |
| 3DMark Time Spy | 36,971 points | jedi95, 2023 |
| 3DMark Time Spy Extreme | 18,335 points | jedi95, 2023; the supplied result does not specify whether this is a total or graphics score |
| 3DMark Port Royal | 20,492 points | jedi95, 2023 |
| 3DMark Speedway | 7,690 points | jedi95, 2023 |
| Peak GPU draw | 696 W | jedi95’s Time Spy Extreme run, 2023 |
How close was it to an RTX 4090?
The strongest like-for-like claim in the available comparisons is for a Time Spy Extreme graphics score: the modified Radeon scored 19,137, against 19,540 for PC Gamer’s RTX 4090 Founders Edition result reported in 2023. That is close performance in this particular raster-oriented benchmark, not proof that the two cards perform alike across games or workloads.
Keep that comparison separate from the 18,335 Time Spy Extreme result listed above: the figures refer to different score descriptions, and the 18,335 result is not identified as a graphics subscore. Nor should the Radeon’s Port Royal result be treated as evidence of ray-tracing parity. HotHardware’s 2023 comparison found the RTX 4090 40% ahead of the hot-clocked Radeon in that ray-tracing test.
Recommended Free Tools
#1 Best Overall
- Flagship Gaming Performance, AMD Radeon RX 7900 XTX GPU with 2615 MHz boost clock and 24GB GDDR6 memory for elite 4K gaming
- Advanced RDNA 3 Architecture, 96 compute units with RT+AI accelerators and 96MB AMD Infinity Cache technology
- Premium Cooling Solution, Phantom Gaming 3X Cooling System with Striped Ring Fans and reinforced metal frame
- High-Speed Memory, 24GB GDDR6 on 384-bit memory bus delivers exceptional bandwidth for 4K gaming and content creation
- Silent Operation, 0dB Silent Cooling technology ensures zero fan noise during low-intensity tasks
Did the modification improve game performance?
HotHardware reported that the modified card was 17% faster in Warzone and 16% faster in Cyberpunk 2077 than the same RX 7900 XTX running at stock settings with water cooling. Those are gains over the Radeon’s own stock configuration, not percentage advantages over an RTX 4090. The experiment therefore shows meaningful but much smaller game uplifts than a claim of universal benchmark parity might suggest.
What did removing the limit actually involve?
The phrase “removes the power limit” can sound like a routine setting change. Here it describes a hardware-level control bypass: the EVC2SE was connected to the card’s I2C bus so the VRM controller’s reported power could be adjusted. Tom’s Hardware characterized the technique as “shunt-like” because it changes the power value reported to the GPU. It does not make the card more efficient or eliminate the electrical and thermal cost of higher power.
The test used an Asus TUF RX 7900 XTX fitted with an EK Quantum Vector full-cover water block and coolant at 10°C. The listed test system included an Intel Core i9-13900KS, DDR5-7400 memory, an Asus Z790 Apex motherboard, Windows 10 22H2, and AMD driver 23.4.3. The results belong to this particular setup and its 2023 software environment.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why is this not a practical gaming setup?
The RX 7900 XTX’s advertised boost clock was 2.5 GHz. Its standard power rating was 355 W, and HotHardware described a software-adjustable limit of about 430 W. Against that context, the measured 696 W peak in Time Spy Extreme shows the scale of the trade-off: the experiment more than doubled the standard board-power figure while relying on chilled liquid cooling and specialist hardware. ComputerBase described the nearly 700 W arrangement as impractical for ordinary gaming.
- Cooling: The 10°C coolant and full-cover block are far removed from a typical air-cooled gaming PC.
- Equipment and modification: The EVC2SE connection and power-reporting adjustment require specialist hardware and work on the graphics card’s control circuitry; this is not a standard driver toggle.
- Evidence scope: This is one enthusiast experiment, not a multi-card study. It does not establish that other RX 7900 XTX cards will reach the same clocks or scores.
For most buyers, these results are best understood as an illustration of frequency headroom under extreme conditions, not as a realistic estimate of the performance of a retail RX 7900 XTX. A compatible full-cover water block may be a relevant accessory for a custom-loop build, but compatibility must be checked for the exact card model.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
