NVIDIA Lite Hash Rate GPU Performance

Always look at the date when you read an article. Some of the content in this article is most likely out of date, as it was written on October 26, 2021. For newer information, see our more recent articles.

Table of Contents

Introduction

In an attempt to make their GeForce line of consumer video cards less appealing to cryptocurrency miners, NVIDIA has updated many of their GPUs with "lite hash rate" (LHR) versions. Graphics cards with LHR are supposedly able to detect when they are being used for crypto mining usage and then reduce their hash performance – in particular, the Dagger Hashimoto algorithm. NVIDIA claims that should reduce the effectiveness of these cards for mining Etherium by about 50%, without impacting game performance or other applications. The reduction to mining speed is of little concern to us here at Puget Systems, but to be sure this change won't impact our workstation users we put a pair of GeForce RTX 3070 cards – one with LHR and one without – to the test in several professional benchmarks.

Test Setup

We used the following platform and software to test the relative performance of GeForce RTX 3070 cards with and without LHR:

Test Platform
CPU	Intel Core i9 11900K
CPU Cooler	Noctua NH-U12S
Motherboard	Gigabyte Z490 Vision D
RAM	4x DDR4-3200 16GB (64GB total)
Video Card	NVIDIA GeForce RTX 3070 8GB NVIDIA GeForce RTX 3070 8GB LHR
Hard Drive	Samsung 980 Pro 2TB
Software	Windows 11 Pro 64-bit GeForce Studio Driver 472.12 Premiere Pro 15.1.0 V-Ray 5 Benchmark 5.01.00 OctaneBench 2020.1.5 Redshift 3.0.57 Metashape Professional 1.7.2 Pix4Dmapper 4.6.4

While we normally use NVIDIA Founders Edition video cards when possible, for this article we instead used EVGA's GeForce RTX 3070 XC3 ULTRA GAMING models. That is because, so far as I am aware, NVIDIA has not made a lite hash rate version of the RTX 3070 Founders Edition – and I wanted to keep everything else about the cards as similar as possible, to avoid something like boost clocks or cooler effectiveness from skewing the results. Here are EVGA's own specs pages showing that these cards should be identical in all respects other than LHR:

Methodology

To capture GPU performance across multiple workflows, we ran tests on three primary types of applications: video editing, rendering, and photogrammetry. This isn't a complete picture, of course, since there are thousands of different software titles which make use of the video card in some form or fashion – but these examples cover a wide range of usage, especially with regards to the work that many of our customers are doing in the field. Each test was run twice, and the better of the two results is included in our charts below.

It should be noted that gaming performance is not being included here, since that is not something we focus on and plenty of other review websites cover that in greater depth than we ever could.

Results

Here are charts showing the performance of the new lite hash rate RTX 3070, in green, versus the older non-LHR card in blue:

NVIDIA GeForce RTX 3070 Normal vs LHR Performance in Premiere Pro

GeForce LHR Performance in Video Editing

GeForce LHR Performance in Rendering

GeForce LHR Performance in Photogrammetry

Conclusion

It is pretty clear from the results above that the GeForce RTX 3070 "lite hash rate" video card did not suffer performance loss in any of the applications we tested. In fact, if anything the LHR variant we tested appears to be slightly faster than the original version – despite being from the same board manufacturer, running on the same driver, using the same cooling system, and having the same base and boost clock speeds. There is variance in GPUs from one individual sample to another, which may be all that we are seeing here – or it could be that the addition of LHR or some other small tweak made in the process actually boosted performance just a tiny bit. To really answer that we'd have to get in additional samples of both LHR and non-LHR cards, ideally from other manufacturers and other parts of the RTX 30 Series… but given how small the difference here is, and that it appears to benefit the newer model, that would likely be overkill. Instead, I am happy to call LHR cards a success in terms of workstation performance!