1usmus - Months of HYDRA research. Then mVolt+ copied the features.

For months, I invested my time and substantial personal funds in NVIDIA GPU research: hardware, test equipment, driver analysis, failed experiments, exact readback and recovery checks. HYDRA’s RTX 40/50 controls did not come from a ready-made tutorial. These breakthroughs cost far more than time alone. My dated records and released builds show when those results became working features.

On 17 August, I publicly demonstrated HYDRA’s RTX 50 VRAM and board Power Limit unlocks, including hardware results. The implementation details remained in our private research, but the result and who produced it were public more than five weeks before the exact mVolt+ Extreme driver-bearing file we can independently verify. HYDRA shipped these controls to Early Access members on 27 August.

I am not going to turn this into a personal shouting match. But I will not stay quiet while months of costly research are treated as a free feature roadmap for another product. After HYDRA demonstrated and released difficult NVIDIA controls, mVolt+ added strikingly similar capabilities in sequence, within days or weeks. Some appeared in a separate Extreme build rather than the ordinary release. I regard this as repeated appropriation of HYDRA’s work, and I am putting the dated record in public.

all times UTC. Each line compares the first verified HYDRA release with the first mVolt artifact we could verify. The intervals measure availability, not the time either developer spent researching a feature.

XBAR / SYS / Video V/F curves: HYDRA build 2313 — 27 August, 17:49; mVolt+ v0.47.1 — 20 September, 17:29. 23 days, 23 hours, 40 minutes later. This compares editable curves, not a basic XBAR clock offset.

XBAR V/F curves across the desktop GeForce RTX 40 series: HYDRA build 2315 — 30 August, 18:24; no working RTX 40 XBAR curve editor verified in the inspected mVolt+ releases through 29 September. Not copied yet.

Exact boost-limit winner for clock domains and voltage rails: HYDRA build 2317 — 2 September, 15:52; the inspected mVolt+ v0.39 binary — 9 September, 18:35. 7 days, 2 hours, 42 minutes later.

Hardware fan curves: HYDRA build 2319 — 9 September, 16:28; mVolt+ v0.39 added a software fan curve — 9 September, 18:35. The visible idea appeared 2 hours, 7 minutes later, but mVolt+ did not reproduce HYDRA’s hardware fan-policy implementation in that build.

RTX 50 board Power Limit ceiling unlock: HYDRA build 2313 — 27 August, 17:49; the first independently verified public mirror of the separate Extreme 0.47.2 executable, containing an embedded driver and a power-policy write path — 23 September, 19:59. 27 days, 2 hours, 9 minutes later to that exact mirrored file.

MCLK ceiling unlock beyond the normal +3000 range: HYDRA build 2313 — 27 August, 17:49; the same verified Extreme 0.47.2 file, with a two-stage MCLK limit-change path — 23 September, 19:59. 27 days, 2 hours, 9 minutes later to that exact mirrored file.

Board Power Limit ceiling unlock across the desktop GeForce RTX 40 series: HYDRA build 2315 — 30 August, 18:24; no working RTX 40 above-VBIOS board-ceiling unlock verified in the inspected mVolt+ builds through 29 September. Not copied yet.

This is a conservative selection of major features. I have deliberately left smaller instances of repeated feature overlap out of the timeline.

The Extreme date comes from a public forum attachment whose SHA-256 exactly matches the file we analyzed: `6D3F1276CCB04EA7DE71EDBF353E18356772ED1C16998A7B0E3F183F44B3F2ED`. It is a third-party mirror, not an official GitHub release date. We also checked the current executables from all 20 official mVolt+ releases: none contains that embedded driver. I will not claim to know when the Extreme author first built it.

One similar feature can be an independent discovery. A succession of unusually specific features following our public work is a different pattern. I do not accept “coincidence” as an adequate explanation for that pattern. This timeline does not prove access to HYDRA’s source code; it does support my conclusion that HYDRA’s research and product work have been repeatedly copied without credit.

Reverse engineering is not the whole story. Some information relevant to our research was available under NDA and cannot be obtained by reversing a publicly distributed driver. I will not disclose that information here. This is not a claim that the mVolt+ author had access to it. The public case in this post rests on the dated HYDRA releases and the inspected mVolt+ artifacts.

I am not waiting for an explanation from the mVolt+ author. The purpose of this post is to put the dated releases and inspected binaries in front of users and the press. HYDRA’s public demonstration, the later mVolt+ artifacts and the rejection of the verified Ada XBAR curve format by the inspected mVolt binaries are independently checkable facts. An RTX 40 board Power Limit ceiling unlock remains unverified in those builds.

To journalists and users: examine the dates and compare the actual functions, not similar names. Any newsroom that wants to verify the claims can test the relevant builds directly. I am publishing this timeline so everyone can judge it. This is not a call to harass anyone.

HYDRA’s 17 August RTX 50 announcement · HYDRA build 2315 / RTX 40 post · HYDRA build 2317 post · HYDRA build 2319 post · mVolt+ releases · Extreme mirror.