RTX 3070 Performance

Performance Summary Charts & Graphs

Gaming Performance Summary Charts

Here are the summary charts of 34 games and 3 synthetic tests. The highest settings were always chosen and the settings are listed on the chart.  The benches were run at 1920×1080, 2560×1440 and at 3840×2160.  Eight cards were compared and they are listed in order starting with the least two powerful cards on the left to the most powerful on the right: the RX 5700 XT, the GTX 1080 Ti, the RTX 2070 SUPER, the RTX 2080 SUPER, the RTX 3070, the RTX 2080 Ti, the RTX 3080, and the RTX 3090.

Most results, except for  synthetic scores, show average framerates, and higher is better. Minimum framerates are next to the averages in italics and in a slightly smaller font.  Games benched with OCAT show average framerates, but the minimums are expressed by frametimes (99th-percentile) in ms where lower are better.  An “X” means the benchmark was not run (or could not be run).

All of the games that we tested ran well except for A Total War Saga: Troy, and we suspect that it still may be a game or driver issue that has not yet been addressed by NVIDIA’s driver team.  Control also had issues with setting the render resolution for 2560×1440 for some cards, and the RX 5700 XT refused to run in DX12 at all.  The Shadow of the Tomb Raider benchmark refused to run on the GTX 1080 Ti and the RX 5700 XT, and it would crash to desktop when we attempted to access the benchmark.  We note that cards with less than 11GB of vRAM cannot run Ghost Recon: Breakpoint at 4K/Ultimate above a slideshow.

Although there is some variability depending on which games were tested, the RTX 3070 FE generally trades blows with the slightly faster RTX 2080 Ti FE.  NVIDIA claims that the RTX 3070 and the RTX 2080 Ti have similar performance, but no doubt they are referring to the entry-level non-overclocked Ti cards – the RTX 2080 Ti Founders Edition (FE) is an overclocked card.

Next we look at ten RTX/DLSS enabled games, each using maximum ray traced settings and the highest quality DLSS where available.

RTX/DLSS Benchmarks

The RTX 3070 is next compared with the same cards when ray tracing or RTX/DLSS are enabled.  The RX 5700 XT can only run one ray traced non-RTX game – Crysis Remastered – so we did not include it.  However, we included the GTX 1080 Ti results even though it cannot run RTX features above a crawl above 1080P in a very few selected games.  We also added RTX Quake II, a fully path-traced game, and the ‘Bright Memory Infinite’ RTX demo benchmark.

The RTX 3070 now appears to be slightly faster than the RTX 2080 Ti when DLSS or RTX features are enabled.  NVIDIA is expecting quite a few RTX-enabled games this year and next year, and Watch Dogs: Legion launches in two days with RTX features.

Next we look at overclocked performance.

Overclocked benchmarks

These benchmarks are run with the RTX 3070 overclocked +100MHz on the core and +100MHz on the memory versus at stock clocks.  The RTX 3070 stock results are presented first and the overclocked RTX 3070 is in the second column.

There is a small performance increase from overclocking the RTX 3070.  NVIDIA has evidently locked Ampere cards overclocking down in an attempt to maximize performance for all Founders Edition gamers.  We would also suggest that the RTX 3070 is rather voltage constrained and if you want a higher overclock, pick a factory-overclocked partner version instead of a Founders Edition.

Let’s look at non-gaming applications next to see if the RTX 3070 is a good upgrade from the other video cards we test starting with Blender.

Blender 2.90 Benchmark

Blender is a very popular open source 3D content creation suite. It supports every aspect of 3D development with a complete range of tools for professional 3D creation.

We benchmarked three Blender 2.90 benchmarks which measure GPU performance by timing how long it takes to render production files. We tested seven of our comparison cards with both CUDA and Optix running on the GPU instead of using the CPU.  We did not benchmark the RX 5700 XT using OpenCL.

For the following chart, lower is better as the benchmark renders a scene multiple times and gives the results in minutes and seconds.

Blender’s benchmark performance is similar using the RTX 3070 compared with the RTX 2080 Ti.

Next we look at the OctaneBench.

OctaneBench

OctaneBench allows you to benchmark your GPU using OctaneRender.  The hardware and software requirements to run OctaneBench are the same as for OctaneRender Standalone.

We run OctaneBench 2020.1.5 for Windows and here are the RTX 3070’s complete results and overall score of 411.73.

We compare with the score and results for the RTX 2080 Ti with 384.24.

Here is the summary chart comparing the TITAN Xp, the RTX 2070 SUPER, the RTX 2080 Ti, the RTX 3070, the RTX 3080, and the RTX 3090.

The RTX 3070 is a very decent card when used for rendering.

Next, we move on to AIDA64 GPGPU benchmarks.

AIDA64 v6.25

AIDA64 is an important industry tool for benchmarkers.   Its GPGPU benchmarks measure performance and give scores to compare against other popular video cards.

AIDA64’s benchmark code methods are written in Assembly language, and they are well-optimized for every popular AMD, Intel, NVIDIA and VIA processor by utilizing the appropriate instruction set extensions.  We use the Engineer’s full version of AIDA64 courtesy of FinalWire.  AIDA64 is free to to try and use for 30 days.  CPU results are also shown for comparison with both the RTX 3070 and GTX 2080 Ti GPGPU benchmarks.

Here are the RTX 3070 AIDA64 GPGPU results compared with an overclocked i9-10900K.

Here is the chart summary of the AIDA64 GPGPU benchmarks with seven of our competing cards side-by-side.

The RTX 3070 is a fast GPGPU card and it compares favorably with the RTX 2080 Ti.  So let’s look at Sandra 2020 next.

SiSoft Sandra 2020

To see where the CPU, GPU, and motherboard performance results differ, there is no better tool than SiSoft’s Sandra 2020.  SiSoftware SANDRA (the System ANalyser, Diagnostic and Reporting Assistant) is a excellent information & diagnostic utility in a complete package.  It is able to provide all the information about your hardware, software, and other devices for diagnosis and for benchmarking.  Sandra is derived from a Greek name that implies “defender” or “helper”.

There are several versions of Sandra, including a free version of Sandra Lite that anyone can download and use.  Sandra 2020 R10 is the latest version, and we are using the full engineer suite courtesy of SiSoft.  Sandra 2020 features continuous multiple monthly incremental improvements over earlier versions of Sandra.  It will benchmark and analyze all of the important PC subsystems and even rank your PC while giving recommendations for improvement.

The author of Sandra 2020 informed us that while NVIDIA has sent some optimizations, they are generic for all cards, not Ampere specific.  The tensors for FP64 & TF32 have not been enabled in Sandra 2020 so GEMM & convolution running on tensors will get faster using Ampere’s tensor cores.  BF16 is supposed to be faster than FP16/half-float, but since precision losses are unknown it has not yet been enabled either.  And finally, once the updated CUDA SDK for Ampere gets publicly released, Sandra GPGPU performance should improve also.

With the above in mind, we ran Sandra’s intensive GPGPU benchmarks and charted the results summarizing them.

In Sandra GPGPU benchmarks, the RTX 3070 is similar in performance to the RTX 2080 Ti.  However, since the architectures are different, each card exhibits different characteristics with different strengths and weaknesses.

SPECworkstation3 (3.0.4) Benchmarks

All the SPECworkstation3 benchmarks are based on professional applications, most of which are in the CAD/CAM or media and entertainment fields. All of these benchmarks are free except for vendors of computer-related products and/or services.

The most comprehensive workstation benchmark is SPECworkstation3.  It’s a free-standing benchmark which does not require ancillary software. It measures GPU, CPU, storage and all other major aspects of workstation performance based on actual applications and representative workloads.  We only tested the GPU-related workstation performance as checked in the image above.

Here are our raw SPECworkstation 3.0.4.summary and raw scores for the RTX 3070.

Here are the SPECworkstation3 results summarized in a chart along with six competing cards.  Higher is better.

Using SPEC benchmarks, the RTX 3070 is similar in performance to the RTX 2080 Ti.  However, since the architectures are different, the cards each exhibit different characteristics with different strengths and weaknesses.

SPECviewperf 2020 GPU Benches

The SPEC Graphics Performance Characterization Group (SPECgpc) has released a new 2020 version of its SPECviewperf benchmark twelve days ago that features updated viewsets, new models, support for both 2K and 4K display resolutions, and improved set-up and results management.

We benchmarked at 4K and here are the summary and the raw results for the RTX 3070.

Here are SPECviewperf 2020 GPU benchmarks summarized in a chart together with the three other cards that we had time to test, the RTX 2080 Ti, the RTX 2070 SUPER and the RX 5700 XT.

The RTX 2080 Ti appears to be a little faster than the RTX 3070 in SPECviewperf 2020 GPU benches, but both cards are generally faster than the RX 5700 XT or the RTX 2070 SUPER.

After seeing these benches, some creative users will probably upgrade their existing systems with a new RTX 30X0 series card based on the performance increases and the associated increases in productivity that they require. The question to buy an RTX 3070 should be based on the workflow and requirements of each user as well as their budget.  Time is money depending on how these apps are used.  However, the target demographic for the RTX 3070 is primarily gaming for gamers, especially at 1440P.

Let’s head to our conclusion.