What GPU is recommended for hashcat password cracking with the best performance?

2024-04-15 10:23:47

With the launch of the 40 series NVIDIA GPU super update, the company's product lineup has become quite complex. In the 4070 series alone, there are four models of NVIDIA GeForce RTX to choose from: the original 4070 and 4070 Ti, as well as the current 4070 Super and 4070 Ti Super. Understanding the differences between these cards and which models provide the best cost-effectiveness in password recovery work is a key consideration for IT professionals.

Why use GPU to crack passwords?

在某些任務中，當今 GPU 的大規模并行計算單元提供的性能遠遠超過計算機的 CPU，同時保持相似的功耗。我們的產品支持硬件加速技術，利用現代 GPU 的可用計算能力來加速密碼恢復攻擊。我們的工具支持使用 AMD、Intel 和 NVIDIA 芯片構建的 GPU，但我們始終推薦 NVIDIA 顯卡，因為它們在 24×7 環境中具有最大的穩定性和可靠性。

我們的加速技術將一些復雜的計算卸載到快速且可擴展的計算單元（NVIDIA CUDA 內核）上，與中央處理器相比，密碼恢復速度提高了 25 到 250 倍。

NVIDIA RTX 40 Series

The first batch of GPUs released in the Ada Lovelace driven 40 series were the powerful NVIDIA GeForce RTX 4090 and 4080 motherboards released at the end of 2022. Following closely behind are models with lower performance. The initial lineup includes RTX 4090, 4080, 4070 Ti, 4070, 4060 Ti (8GB and 16GB versions), and 4060.

We benchmarked some of the GPUs in the NVIDIA RTX 40 series graphics cards: a faster and more efficient password recovery accelerator, and concluded that the mid-range NVIDIA GeForce RTX 4070 Ti is faster and has lower power consumption than the previous generation champion RTX 3090. Compared to the older but similarly priced RTX 3070 Ti motherboard, the new RTX 4070 has improved performance by about 40% and reduced power consumption by about 30%. This provides a strong basis for recommending the Ada Lovelace generation to perform password recovery tasks instead of Ampere.

In January 2024, NVIDIA updated its lineup by adding several models labeled "Super". The "Super" version includes 4080 Super, 4070 Ti Super, and 4070 Super; The low-end model GPU has not received an update. NVIDIA subsequently discontinued some old models (4080 and 4070 Ti), although they can still be sold. At the same time, the 4070 and 4070 Super cards remain a part of the company's current product line, while the RTX 4070 is being discounted. The new lineup is not very simple, especially considering that discontinued models are still being sold together with updated cards. It is important to understand the diversity of GPU models and the performance they provide when selecting a GPU for password recovery tasks.

For password recovery, the number of VRAMs is largely irrelevant. Under the same conditions, there is no difference in pure attack speed between the NVIDIA GeForce RTX 4060 Ti (8GB) and GeForce RTX 4060 Ti (16GB) variants, but the latter has a higher cost than the former. However, compared to the discontinued 12GB 4070 Ti, the 16GB 4070 Ti Super not only doubles the number of VRAMs, but also expands the interface width from 192 bits to 256 bits, while increasing the number of execution units. Therefore, the "super" version of 4070 Ti has improved performance by 10% in dual precision calculations.

When it comes to performance, people can benchmark graphics cards in many different ways and present results in different units. We are unable to test all GPU models in our settings, so for comparison between Apple and Apple, we rely on the data in the specification sheet and data obtained from independent sources. Specification list numbers for single precision, double precision, semi precision, FP16, and TFLOPS; For password recovery tasks, only the double precision metric is relevant. Comparing different models based solely on this metric (and importantly, within the same architecture) allows us to predict their relative performance quite accurately. With this information, you can evaluate their performance in certain tasks; We suggest using the numbers from the test as a basis.

Note: Other test data can be found throughout the entire article.

Our cost-effectiveness data is only based on the suggested retail price in the United States. Street prices may and do vary; Remember this when purchasing the most cost-effective circuit board. Finally, make sure to understand the impact of TDP on the circuit board in order to build an efficient, well cooled, and adequately powered setup.

Note: The above specifications are for reference only and may vary depending on OEM and specific models.

Based on this information, we can draw several conclusions:

The "super" update of RTX 4070 has brought significant performance improvements, with only a slight increase in power consumption. Although the original RTX 4070 remains in NVIDIA's lineup, we cannot recommend it even after a price reduction.

The "super" version of RTX 4080 did not bring significant performance improvements, but it greatly improved cost-effectiveness, mainly due to the price reduction of updated models. This in turn means that if offered at the same or lower price as the new Super variant, a non Super version of the RTX 4080 can be considered.

The cost-effective RTX 4070 Super model demonstrates excellent cost-effectiveness, but only when purchased at the suggested retail price.

The update of NVIDIA's 40 series super GPU involves considering factors such as performance indicators, pricing changes, and TDP. By understanding these factors, IT professionals can make informed decisions about which GPU models are most suitable for their specific requirements and budget.