{"id":893834,"date":"2026-10-25T06:14:14","date_gmt":"2026-10-25T06:14:14","guid":{"rendered":"https:\/\/www.newsbeep.com\/us\/893834\/"},"modified":"2026-10-25T06:14:14","modified_gmt":"2026-10-25T06:14:14","slug":"nvidia-rtx-30-rtx-40-gpus-get-impressively-fast-in-ai-llms-as-compute-doubles-with-new-unlock-tool-but-prices-surge-too","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/us\/893834\/","title":{"rendered":"NVIDIA RTX 30 & RTX 40 GPUs Get Impressively Fast In AI LLMs As Compute Doubles With New Unlock Tool But Prices Surge Too"},"content":{"rendered":"

A new GPU unlock tool<\/a> has unlocked the full potential of NVIDIA’s RTX 30 & RTX 40 GPUs, achieving faster AI LLM performance and compute capabilities.<\/p>\n

Back in the crypto mining days, NVIDIA had the compute performance of its GeForce RTX 30 GPUs virtually unlocked through BIOS firmware and patches as it rolled out its mining-specific CMP line of chips. The CMP cards have also recieved a special unlock tool<\/a> which exposes their full memory capabilities, jumping from 8-10 GB to 64-80 GB.<\/p>\n

While modders managed to unlock the full mining rates, the compute restrictions were still in place. But that changes through a new open-source GitHub project called “Zagita<\/a>,” which is unlocking the full BF16\/FP16 rates featured on the Ampere GPUs<\/a>. The tool is easy to use and doesn’t require any vBIOS flash or hardware modding.<\/p>\n

\ud83d\udea8 \u4f60\u7684 30 \u7cfb\u663e\u5361\uff0c\u7b97\u529b\u88ab\u5b98\u65b9\u9501\u4e86\u4e00\u534a<\/p>\n

\u4e00\u4e2a\u5185\u6838\u6a21\u5757\u5728 nvidia \u9a71\u52a8\u7ed1\u5b9a\u4e4b\u524d\u52a8\u624b\uff0c\u628a FP\/BF16 \u6307\u4ee4\u7684\u53d1\u5c04\u901f\u7387\u653e\u5230\u4e13\u4e1a\u5361\u7684\u6c34\u5e73\u3002\u4e0d\u662f\u5237 vBIOS\uff0c\u4e5f\u4e0d\u662f\u6539\u9a71\u52a8\uff0c\u9a71\u52a8\u6bcf\u6b21\u66f4\u65b0\u90fd\u4e0d\u7528\u91cd\u6253\u8865\u4e01\u3002<\/p>\n

1\ufe0f\u20e3 3090 Ti \u5b9e\u6d4b\uff1aBF16 \u4ece 87.5 \u6da8\u5230 157 TFLOPS\uff0cTF32 \u4ece 42.6 \u6da8\u5230 82.6
2\ufe0f\u20e3 3080 10GB \u5728\u539f\u5382 320W\u2026<\/p>\n

\u2014 MrWhite \u767d\u5148\u751f (@neuronxwhite) October 24, 2026<\/a><\/p>\n

Running the tool is simple, and it unlocks double the compute capabilities of NVIDIA’s RTX 30 series FP16\/BF16 FMA across all GA102, GA103, GA104, and GA106 GPUs.<\/p>\n

Unlock FP\/BF16 instruction issue rate on Nvidia 3000’s cards to their professional equivalents<\/p>\n

This Linux kernel module combines the CMP unlocker’s V67\/PLM technique with the pre-driver timing of the UEFI approach. It runs the unlock, writes the speed selectors (SS0 0x88888888, which selects the full BF16 rate, and the 4-bit SS1 0x8 IMLA4) and performs FLR before the unmodified NVIDIA driver binds the card. It is neither a UEFI application nor a patched NVIDIA driver.<\/p>\n

The benefit of this early kmod driver is that it does not require manual patching each time nvidia releases a new driver version.<\/p>\n

The effect of this unlock is that it doubles the instruction issuing speed of FP\/BF16 FMA into FP32 operations on GA102, GA103, GA104 and GA106 cards.<\/p>\n

Many users are running the tool now and seeing a boost of up to 80% on the RTX 30 series GPUs. For example, the RTX 3090 Ti has its BF16 capabilities boosted from 87.5 TFLOPs to 157 TFLOPs & TF32 goes from 42.6 TFLOPs to 82.6 TFLOPs. Similarly, the RTX 3080 10 GB goes from 65 TFLOPs of FP16 to 113 TFLOPs at its stock 320W power.<\/p>\n

Power limitFP16BF16per MHzBefore300 W59.6 TFLOPS @ 1384 MHz63.0 @ 14620.043After, cold boot380 W100.3 @ 1208105.9 @ 12690.083After, warm reboot350 W89.5 @ 107295.1 @ 11480.083<\/p>\n

3090 Ti dense 8192\u00b3 GEMM at 450 W, five-sample medians with FP32 accumulation:<\/p>\n

InputBeforeAfterFP1687.46 TFLOPS118.39 TFLOPSBF1687.50 TFLOPS157.41 TFLOPSTF3242.58 TFLOPS82.61 TFLOPS<\/p>\n

In AI LLMs, an unlocked RTX 30 GPU was able to offer 33-36% faster Prefill (long-context) output, while the latency in TTFT (Time-to-first-token) was reduced by 25%. Decode performance wasn’t affected, and matrix-multiply performance remained identical.<\/p>\n

ChipCardPCI IDResultBF16 GEMM gainGA102RTX 3090 Ti10de:2203Unlocked+68\u201380%GA102RTX 309010de:2204Unlocked+25\u201367%GA102RTX 3080 Ti 20 GB (OEM, GA102-250)10de:2205Unlock completes, BF16 stays at half rate\u2014GA102RTX 3080 10 GB10de:2206Unlocked+70\u201373%GA102RTX 3080 20 GB (memory upgrade)10de:2206Unlocked+56\u201359%GA104RTX 307010de:2484Unlocked+73\u201381%GA104RTX 3070 Ti10de:2482Not tested with this module (the same unlock worked from a UEFI port on Windows)\u2014GA106RTX 3060 12 GB10de:2504Unlockedabout +80%GA103any10de:2414No reports yet\u2014<\/p>\n

Currently, the unlock is only available on Linux platforms; Windows isn’t supported yet. The RTX 30 GPU lineup sees the biggest benefit, but the unlock does extend to older GTX 9, GTX 10, & RTX 20 series cards.<\/p>\n

Not just 30 Series but 40 Series GPUs too seemingly see their Tensor Core throughput doubled. I believe this might end up increasing these GPUs’ 2nd-hand prices as well. It’ll also be interesting to see if this can be leveraged to improve DLSS perf since it runs on Tensor Cores. https:\/\/t.co\/8eN6TUtgPW<\/a><\/p>\n

\u2014 Sebastian Castellanos (@Sebasti66855537) October 24, 2026<\/a><\/p>\n

Even RTX 40 GPUs are seeing an uplift with these new unlocks.<\/p>\n

\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t<\/a><\/p>\n

Other users are also fine-tuning models on older RTX 30 GPU cards such as the 24 GB RTX 3090. A user has reportedly reached 381 tokens per second throughput on Qwen3.8-27B.<\/p>\n

Qwen3.8-27B reportedly reached 381 tokens\/sec on a single RTX 3090 with 24GB of VRAM.<\/p>\n

The starting point was 82 tokens\/sec. That is approximately a 4.6\u00d7 increase without changing the GPU or the model.<\/p>\n

Developer TeksEdge achieved this by optimizing how the model generates,\u2026<\/p>\n

\u2014 FHILY\ud83d\udc51 (@Oluwaphilemon1) October 24, 2026<\/a><\/p>\n

The AI model ran at a baseline performance of 82 TPS and improved to a 4.6x rate through optimized MTP, DFlash2 with context lookup, and an optimized lookup verification. We reported this story this week<\/a>, and as models continue to be optimized on older cards, their prices are also seeing an uptick.<\/p>\n

\ud83d\udd25\u725b\u903c\u4e86\uff0112GB\u663e\u5361\u672c\u5730\u8dd1763B\uff01DeepSeek-V4.1-Flash-1017\u91cf\u5316\u7248\u70b8\u88c2\u767b\u573a\uff0c\u524d\u6cbf\u8d28\u91cf\u51e0\u4e4e\u96f6\u635f\u5931\uff01<\/p>\n

ISTA-DASLab\u56e2\u961f\u53d1\u5e03GSQ-RCO\u975e\u5747\u5300\u91cf\u5316\u7248\uff0cStrata\u5f15\u64ce\u539f\u751f\u652f\u6301\u3002763B\u53c2\u6570MoE\u5927\u6a21\u578b\u6700\u4f4e\u538b\u52301.75 bpw\uff0c\u5728\u6d88\u8d39\u7ea7\u786c\u4ef6\u4e0a\u5c31\u80fd\u8dd1\uff01<\/p>\n

\u6838\u5fc3\u4eae\u70b9
\ud83d\udd39\u6781\u81f4\u538b\u7f29 + \u51e0\u4e4e\u65e0\u635f\uff1aIQ2_XXS\uff082.00\u2026 https:\/\/t.co\/yWTvTW8LZz<\/a> pic.twitter.com\/eAiRxXcBIN<\/a><\/p>\n

\u2014 SuSu_\u9165\u9165\ud83d\udc45 (@NFT_Chen) October 24, 2026<\/a><\/p>\n

For example, the NVIDIA GeForce RTX 3090 24 GB was selling for around $1000 US, but since these reports, the prices have climbed past $2000.<\/p>\n

\"An<\/a><\/p>\n

The RTX 3090 Ti is fast approaching the $ 3000 US mark, and the RTX 3080 10-12 GB models are reaching $900-$1100 US on eBay<\/a>. The optimizations are giving Ampere graphics cards a new life outside of gaming, but that also means that gamers who were planning to purchase a used RTX 30 card will now have to deal with surging prices due to AI.<\/p>\n

This unlock tool is breathing new life into Ampere-era RTX 30 cards for LLM workloads, delivering substantial compute gains that make them far more viable for inference and fine-tuning without any hardware mods. While the performance uplift is a clear win for AI enthusiasts, the resulting demand is pushing used-market prices sharply higher, turning once-affordable gaming GPUs into sought-after AI accelerators and making them less accessible for pure gamers.<\/p>\n

\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\"Hassan
\n\t\t\t\t\t<\/a><\/p>\n

\n\t\t\t\t\tAbout the author<\/a>: A Software Engineer by training and a PC enthusiast by passion, Hassan Mujtaba serves as Wccftech’s Senior Editor for hardware<\/a> section. With years of experience in the industry, he specializes in deep-dive technical analysis of next-generation CPU and GPU architectures, motherboards, and cooling solutions. His work involves not only breaking news on upcoming technologies but also extensive hands-on reviews and benchmarking.\t\t\t\t<\/p>\n

\n\t\t\t\tFollow Wccftech on Google<\/a> to get more of our news coverage in your feeds.\n\t\t\t<\/p>\n

\t\t