Group check: Nvidia GeForce RTX 40 Collection GPUs

 
Which GeForce RTX 40 Collection GPU provides the most effective bang in your buck for CG work? To seek out out, Jason Lewis places Nvidia’s new client GPUs – and a number of other older playing cards – by way of an exhaustive set of real-world exams.

 
Welcome to a different group check. In by far the most important GPU evaluate I’ve completed for CG Channel, we will probably be looking at three playing cards from the GeForce RTX 40 Collection, Nvidia’s present vary of client GPUs.

Nvidia claims important efficiency enhancements for the RTX 40 Collection GPUs over their predecessors, and lots of exams have proven that that is the case for gaming. However what about digital content material creation?

That’s what this text goals to search out out. We will probably be pitting the GeForce RTX 40 Collection playing cards in opposition to older Nvidia GPUs, together with these from its RTX 30 Collection and RTX 20 Collection, in a collection of real-world DCC exams.
 

Bounce to a different a part of this evaluate
Know-how focus: GPU architectures and APIs
Specs
Testing process
Benchmark outcomes
Different concerns
Verdict

 
Which GPUs are included within the group check?
Because the finish of 2022, Nvidia has been pushing out its GeForce RTX 40 Collection of client GPUs, based mostly on its Ada Lovelace structure.

That is its present era of GPUs, meant to switch the earlier GeForce RTX 30 Collection, based mostly on the Ampere structure, and GeForce RTX 20 Collection, based mostly on the Turing structure.

On this evaluate, we will probably be three playing cards from the GeForce RTX 40 Collection: the high-end GeForce RTX 4090, 4080 and 4070 Ti. (On the time the benchmark testing was completed, the mid-range GeForce RTX 4070 and GeForce RTX 4060 Household had not been launched.)

We will probably be testing them in opposition to two playing cards from the GeForce RTX 30 Collection, the GeForce RTX 3090 and GeForce RTX 3070, and one from the GeForce RTX 20 Collection, the GeForce RTX 2080.

We can even be testing three of Nvidia’s workstation playing cards, the Ampere-generation RTX A6000, and the Turing-generation Titan RTX and Quadro RTX 8000.

That makes the total line-up of GPUs on check as follows:
 

Know-how focus: GPU architectures and APIs

Earlier than I get to the evaluate itself, here’s a fast recap of a number of the technical phrases that you’ll encounter in it. When you’re already aware of them, you could need to skip forward.

Like Nvidia’s previous-generation Ampere and Turing GPUs, the present Ada Lovelace GPU structure options three kinds of processor cores: CUDA cores, designed for rasterization and basic GPU computing; Tensor cores, designed for machine studying operations; and RT cores meant to speed up ray tracing.

With a purpose to benefit from the RT cores, software program has to entry them by way of a graphics API: within the case of the functions featured on this evaluate, both DXR (DirectX Raytracing), or Nvidia’s personal OptiX.

In lots of renderers, the OptiX rendering backend is offered as an alternative choice to an older backend based mostly on Nvidia’s CUDA API. The CUDA backends work with a wider vary of Nvidia GPUs and software program functions, however enabling OptiX normally improves efficiency.

Specs

You’ll find articles on different tech web sites in regards to the Ada Lovelace structure itself, so I gained’t delve into it right here. As a substitute, right here is an outline of the three GeForce RTX 40 Collection GPUs on check.

First, the GeForce RTX 4090. Meant as a direct alternative to the Ampere era’s RTX 3090 and 3090 Ti, the RTX 4090 is at present Nvidia’s quickest client GPU, sporting its top-tier AD102 GPU with 16,384 CUDA cores, 512 fourth-generation Tensor Cores, and 128 third-generation RT cores operating at a core clock velocity of two.23 GHz, and a lift velocity of two.52 GHz. Mix that with 24 GB of 21 GB/s GDDR6X reminiscence on a 384-bit bus, and that provides you 82.6 TFlops of single precision (FP32) compute energy, 1.29 TFlops of double precision (FP64) compute energy, and a reminiscence bandwidth of 1,008 GB/s.

The RTX 4090 I examined is a Founders Version card, and its kind issue is almost an identical to the RTX 3090 Founders Version. Like its predecessor, it’s an enormous card, taking over three complete slots, and weighing in at 77.2 oz (2.2 kg): heavy sufficient that I purchased a GPU brace to carry up the again finish of the cardboard.

At normal clock speeds, the RTX 4090 has a TDP of 450W (Nvidia says that it might eat 600W if pushed past manufacturing facility defaults), and wishes a minimum of a 850W PSU. It makes use of a brand new 16-pin 12VHPWR energy connector, so for anybody with out the newest PSUs, Nvidia provides a 16-pin to 4 x 8-pin PCIe adapter.

On the RTX 4090’s launch, some customers reported points with cables melting, however later analysis from tech web sites urged that the probably trigger was merely person error. Having double- and triple-checked that my very own check card’s connectors had been correctly seated, I’m blissful to report that I’ve not had any issues.

The second new card is the GeForce RTX 4080. Though it’s nonetheless an enthusiast-class GPU, it’s meant to be a extra mainstream card than the RTX 4090, and takes the place of the previous-generation RTX 3080 and 3080 Ti. The RTX 4080 comes outfitted with an AD103 GPU with 9,728 CUDA cores, 304 fourth-generation Tensor cores, and 76 third-generation RT cores. Just like the 4090, it makes use of GDDR6X reminiscence, however its body buffer will get reduce down from 24 GB to 16 GB, and it will get a smaller 256-bit bus.

The RTX 4080 additionally runs barely slower, with a base clock velocity of two.21 GHz, and a lift clock velocity of two.51 GHz. Put all these numbers collectively and also you get 48.7 TFlops of single precision compute energy, 0.76 TFlops of double precision compute energy, and a reminiscence bandwidth of 717 GB/s.

The RTX 4080 on check was additionally a Founders Version card. It’s almost an identical in measurement to the RTX 4090 Founders Version however, at 75.1 oz (2.1 kg), very barely lighter. At manufacturing facility speeds, it has a TDP of 320W and requires a minimum of a 750W PSU. The 4080 makes use of the identical 16-pin 12VHPWR connector because the RTX 4090, the distinction being that the included adapter is a 16-pin to three x 8-pin PCIe mannequin.

The ultimate new card is the GeForce RTX 4070 Ti. It’s an fascinating product, because it was initially as a consequence of be a a second RTX 4080 SKU, however as a consequence of complaints from customers, Nvidia rebranded it and decreased the worth barely.

The RTX 4070 Ti has an AD104 GPU with 7,680 CUDA cores, 240 fourth-generation Tensor Cores, and 60 third-generation RT cores. It additionally makes use of GDDR6X reminiscence and additional shrinks the body buffer to 12GB on a 192-bit bus. The clock speeds for the 4070 Ti are literally larger than both the RTX 4080 and the 4090: the Asus card on check has a base clock velocity of two.31 GHz and a lift clock velocity of two.73 GHz, larger than Nvidia’s reference enhance clock velocity of two.61 GHz. It gives 40.1 TFlops of single precision compute energy, 0.63 TFlops of double precision compute energy, and has a reminiscence bandwidth of 504 GB/s.

One other fascinating side of the 4070 Ti is that there is no such thing as a Founders Version mannequin. Because of this, I examined an Asus TUF Gaming version card. Just like the RTX 4090 and RTX 4080, the Asus RTX 4070 Ti is a three-slot design, though the cooling association could be very completely different from the 2 Founders Version playing cards, with three axial followers pulling air from the identical path, slightly than a two-fan design.

Regardless of the variations in board format, the Asus RTX 4070 TI is roughly the identical measurement because the RTX 4090 and RTX 4080, however considerably lighter: 47.8 oz (1.4 kg). At manufacturing facility speeds, it has a TDP of 285W, with Asus recommending a minimum of a 750W PSU.

Click on the picture to view it full-size.

Testing process

For the check machine, I’m nonetheless utilizing the reliable Xidax AMD Threadripper 3990X system that I reviewed in 2020. Even supposing it’s now three years outdated, it’s nonetheless an especially succesful system and doesn’t look like a bottleneck for any of the GPUs examined.

The present model of the check system has the next specs:

CPU: AMD Threadripper 3990X
Motherboard: MSI Creator TRX40
RAM: 64 GB of three,600 MHz Corsair Dominator DDR4
Storage: 2TB Samsung 970 EVO Plus NVMe SD / 1 TB WD Black NVMe SSD / 4 TB HGST 7,200 rpm HD
PSU: 1300W Seasonic Platinum
OS: Home windows 11 Professional for Workstations

The one GPU not examined on the Threadripper system was the GeForce RTX 3070. I not have entry to a desktop RTX 3070, so testing was completed utilizing the cellular RTX 3070 within the Asus ProArt Studiobook 16 laptop computer from this latest evaluate. In that evaluate, I made up my mind that throughout a variety of exams, the cellular RTX 3070 was round 10% slower than its desktop counterpart, so right here, I added 10% to the scores to approximate the efficiency of a desktop RTX 3070 card. It’s not an excellent methodology, nevertheless it will get us to the correct ballpark.

For testing, I used the next functions:

Viewport efficiency
3ds Max 2023, Blender 3.3.1, Chaos Vantage 1.8.2, D5 Render 2.3.4, Fusion 360, Maya 2023, Modo 16.0v2, Omniverse Create 2022.3.1, SolidWorks 2022, Substance 3D Painter 8.2.0, Unigine Group 2.16.0.1, Unity 2022.1, Unreal Engine 5.1.0, 5.0.3 and 4.27.2

Rendering
Arnold for Maya 5.1.0, Blender 3.3 (Cycles renderer), KeyShot 11.2.0, LuxCoreRender 2.6, Maverick Studio 2022.5, OctaneRender 2022.1 Standalone, Redshift 3.5.12 for 3ds Max, SolidWorks Visualize 2022, V-Ray GPU 6 for 3ds Max Hotfix 3

Different benchmarks
Axiom 3.0.1 for Houdini 19.5, Cinema 4D v2023.1 (Pyro solver), Metashape 1.7.4, Premiere Professional 2022

Artificial benchmarks
3DMark Pace Approach 1.0 and Port Royal 1.2, CryEngine Neon Noir Ray Tracing Benchmark, OctaneBench 2020.1.5, Redshift Benchmark v3.5, Unity Enemies Demo, V-Ray Benchmark v5.02.01

All benchmarking was completed with Nvidia Studio Drivers put in for the GeForce RTX GPUs and Enterprise Drivers put in for the RTX A6000 and Quadro RTX 8000. You’ll find a extra detailed dialogue of the drivers used later within the article.

Within the viewport and modifying benchmarks, the body fee scores characterize the figures attained when manipulating the 3D belongings proven, averaged over 5 testing classes to remove inconsistencies. In all the rendering benchmarks, the CPU was disabled so solely the GPU was used for computing.

This time, testing was completed on a correct productiveness monitor setup, consisting of a pair of 27″ 4K screens operating at 3,840 x 2,160px and a 34” widescreen show operating at 3,440 x 1,440px. All three shows had a refresh fee of 144Hz. When testing viewport efficiency, the software program viewport was constrained to the first show (one of many 27″ screens): no spanning throughout a number of shows was permitted.

Benchmark outcomes

Viewport efficiency
The viewport benchmarks embrace a variety of key DCC functions – general-purpose 3D software program like 3ds Max, Blender and Maya, extra specialist instruments like Substance 3D Painter, CAD packages like SolidWorks and Fusion 360, and real-time 3D functions like D5 Render, Unity and Unreal Engine.

 
As you may count on, the GeForce RTX 4090 takes the lead in virtually all of the viewport exams, with the opposite GeForce RTX 40 Collection GPUs shut behind, however there are few anomalies value mentioning.

First, the CAD packages, SolidWorks and Fusion 360. Within the Solidworks benchmark, each the previous-generation RTX A6000 and the two-generations-old Quadro RTX 8000 beat all the GeForce GPUs on check, suggesting that Solidworks prefers Nvidia’s skilled drivers. Even the Turing-generation Titan RTX comes fairly near the efficiency of the GeForce RTX 4090.

With Fusion 360, the sample is analogous, with the RTX A6000 and Quadro RTX 8000 matching the GeForce RTX 4090 and beating all the different GPUs, though right here, the Titan RTX falls a good distance behind.

The opposite anomaly is Unreal Engine 5.1. I used a number of Unreal scenes for benchmarking. On the 2 most demanding – the town scene and the Valley of the Ancients undertaking – the GeForce RTX 4080 GPUs had the best body charges, however solely by tiny margins. On much less taxing scenes, the margins had been a lot bigger.

Digging deeper, I found that when operating the 2 most complicated scenes, GPU utilization within the Unreal Editor was very low: round 30-40% for the GeForce RTX 40 Collection GPUs; barely larger for the 30 Collection and 20 Collection. CPU and reminiscence utilization and disk I/O had been nowhere near most, so there have been no bottlenecks elsewhere within the system. My suspicion is that this low GPU utilization could also be the results of new options in Unreal Engine 5 – significantly Nanite and World Partition – decreasing pressure on the GPU.

Rendering
Subsequent, we now have a set of GPU rendering benchmarks, carried out with an assortment of the extra fashionable GPU renderers, rendering single frames at 4K or larger resolutions.

 
Within the rendering benchmarks, the GeForce RTX 40 Collection GPUs ship some fairly important efficiency will increase over the previous-generation playing cards, with a number of notable exceptions.

The primary is with the open-source LuxCoreRender, the place the GeForce RTX 3090 and the Titan RTX beat out the GeForce RTX 40 Collection GPUs. That is probably as a consequence of the truth that LuxCoreRender doesn’t get up to date as regularly because the industrial functions and has not been optimised for the brand new GPU {hardware} but.

The second, and extra essential, exception is the V-Ray GPU tank scene. That is designed to emphasize GPU reminiscence, inflicting playing cards with lower than 24 GB of VRAM to go out-of-core: as soon as GPU reminiscence fills up, information have to be transferred between it and system RAM, inflicting important slow-downs.

Right here, the 24 GB GeForce RTX 4090 comes out on high, however a lot of the older playing cards beat the GeForce RTX 4080 and RTX 4070 Ti, merely as a consequence of the truth that in addition they have a lot bigger body buffers.

That is a type of cases the place a case may be made for the skilled GPUs, regardless of their a lot larger price: their bigger body buffers make it potential to render far more complicated scenes with out points.

Lastly, we now have the Arnold scene. Though the GeForce RTX 40 Collection GPUs outperform a lot of the older playing cards, the cellular RTX 3070 within the Asus ProArt laptop computer system beats all of them by a substantial margin. Contemplating that it has the smallest body buffer of all of the GPUs on check, that is very odd. My suspicion is that that is because of the newer Zen 3 structure of the CPU within the laptop computer, in comparison with the Zen 2 structure of the desktop system, however that is nonetheless a reasonably large influence for a CPU to have on a GPU rendering check.

Different benchmarks
The following benchmarks check using the GPU for extra specialist duties. Premiere Professional makes use of the GPU for video encoding; photogrammetry utility Metashape makes use of the GPU for picture processing and 3D mannequin era; and Houdini plugin Axiom and Cinema 4D’s Pyro solver each use the GPU for fluid simulation.

 
Maybe because of the vary of duties they check, the scores for the extra benchmarks are in every single place.

In Premiere Professional, all the new GPUs do very properly, with the RTX 4090 taking the highest spot. The extra fascinating result’s that the RTX 4070 Ti beats the RTX 4080, as do the older RTX 2080 Ti and Quadro RTX 8000 – not by a lot, however a win continues to be a win.

The Metashape outcomes comply with the same sample, in that the RTX 4090 and RTX 4070 Ti take the highest spots, however the relative variations between playing cards are a lot smaller as Metashape doesn’t tax the GPU significantly laborious: common GPU utilization is round 20%.

Subsequent, we now have Cinema 4D’s new Pyro solver. Like GPU rendering, fluid simulations are graphics-memory-intensive, and right here, efficiency is nearly solely dictated by the quantity of GPU reminiscence out there. Throughout testing, I seen that efficiency of all the GPUs was good till I hit their reminiscence limits, after which issues would grind to a near-halt. With 48GB of on-board reminiscence, the RTX A6000 and Quadro RTX 8000 beat all the newer GeForce RTX 40 Collection playing cards by a big margin.

In Axiom’s well-optimised inner benchmark, the outcomes are extra predictable, with the RTX 4090 and RTX 4080 taking the highest spots, adopted by the previous-generation RTX 3090.

Artificial benchmarks
Lastly, we now have an assortment of artificial benchmarks. They don’t precisely predict how a GPU will carry out in manufacturing, however they’re a measure of its efficiency relative to different GPUs, and the scores may be in comparison with these out there on-line for different playing cards.

 
As anticipated, the RTX 40 Collection GPUs take the highest spots right here, by reasonable to important margins. The even scaling in scores is spectacular, however once more, the outcomes aren’t actually consultant of real-world efficiency.

Different concerns

Energy consumption
To check the facility utilization of the GeForce RTX 40 Collection, I measured the facility consumption of all the check system on the wall outlet, utilizing a P3 Kill A Watt meter. Because the check machine is a power-hungry Threadripper system, my figures will probably be larger than most traditional DCC workstations.

For this group check, I measured each energy and present drawn. Present (Amperage) is usually ignored by reviewers, however it may be a essential determinant of what number of machines you may run on a single circuit.

Most US homes run 15A circuits from the principle panel, and lots of circuit breakers are rated for 80% of their most load, so a 15A circuit with an ordinary breaker shouldn’t exceed 12A for steady utilization. In my exams, the present drawn by the check system approached 6A when the extra power-hungry GPUs had been put in. If the wall shops in your house workplace are linked by a single circuit, this might decide whether or not you may run two workstations concurrently, significantly whenever you consider screens and lights.

 
Whereas the GeForce RTX 40 Collection GPUs draw a good quantity of energy, they’re really extra environment friendly than the older Ampere and Turing playing cards in lots of the exams. In my guide, elevated efficiency for little enhance in energy utilization – in some circumstances, decreased energy utilization – is a profitable mixture.

Overclocking
The GeForce RTX 4090 has an ordinary clock velocity of two.23 to 2.52 GHz (in apply, my check card appeared to run at a relentless 2.52 GHz beneath load), at which velocity, its TDP is rated at 450W. However the 16-pin to 4 x 8-pin adapter cable that Nvidia gives can run the RTX 4090 at inventory speeds with solely three of the 4 8-pin PCIe connectors plugged in. Plug within the fourth, and you may push the GPU to a TDP of 600W.

To check how a lot this improves efficiency, I repeated a number of of the benchmarks with the RTX 4090 overclocked to a base CPU clock velocity of three.00 Ghz and a base reminiscence clock velocity of 1,463 MHz.

 
Overclocking the RTX 4090 improves efficiency in a lot of the exams, with GPU rendering, simulation and the artificial benchmarks exhibiting enhancements of 10-20%. There was much less impact on viewport body charges, and in a single case, viewport efficiency really went down.

Though operating temperatures barely elevated, whether or not the efficiency beneficial properties are value the additional energy used is so that you can resolve.

Drivers
Lastly, a be aware on the Studio Drivers with which I benchmarked the GeForce RTX GPUs. Nvidia now provides a alternative of Studio or Sport Prepared Drivers for GeForce playing cards, recommending Studio Drivers for DCC work and Sport Prepared Drivers for gaming. In my exams, I discovered no discernible distinction between them by way of efficiency or show high quality. My understanding is that the Studio Drivers are designed for stability in DCC functions, and whereas I haven’t had any actual points when operating DCC software program on Sport Prepared Drivers, if you’re utilizing your system primarily for content material creation, there is no such thing as a cause to not use the Studio Drivers.

Verdict

From these check outcomes, it’s clear that the GeForce RTX 40 Collection GPUs present a big efficiency enchancment over their predecessors, as long as the duties you need to carry out on them match into graphics reminiscence. When you do a number of memory-intensive work, you could be higher off shopping for a previous-generation card with extra on-board reminiscence: a GeForce RTX 3090 slightly than a GeForce RTX 4080, for instance.

Let’s take a look at the three new playing cards on check one after the other, starting with the GeForce RTX 4090. Other than the worstation-class RTX 6000 Ada Technology – not examined right here – it’s the quickest GPU you should buy proper now, both for gaming or content material creation. Along with the huge efficiency beneficial properties over the previous-generation RTX 3090, it makes use of rather less energy, runs slightly bit cooler, and has room for overclocking.

Its MSRP of $1,599 is a troublesome capsule to swallow, however contemplate it in context. Two generations in the past, the most affordable 24GB Nvidia GPU was the a lot slower Titan RTX, which had a launch worth of $2,499, and earlier than that, the one Nvidia playing cards formally licensed to be used with DCC software program had been the much more costly Quadro GPUs. Though the RTX 4090 prices $100 greater than its quick predecessor, the RTX 3090, it provides almost double the efficiency in some conditions.

Subsequent, we now have the GeForce RTX 4080. That is an fascinating product. It’s a robust GPU, and beats all the previous-generation GeForce RTX 30 Collection GPUs fairly comfortably, besides in exams that require greater than 16 GB of graphics reminiscence.

However with a MSRP of $1,199, the GeForce RTX 4080 is simply too shut in worth to the RTX 4090, provided that it’s a very good deal slower, and comes with considerably much less on-board reminiscence. At $1,000, it might be a better promote. If $1,199 is essentially the most you may spend, the RTX 4080 will serve you properly, however when you can afford it, I’d suggest paying the additional $400 for the RTX 4090.

Lastly, we now have the GeForce RTX 4070 Ti. After I began testing, I didn’t count on it to do properly, so I used to be stunned how strongly it carried out. In lots of the benchmarks, it was nipping on the heels of the RTX 4080, and in a number of circumstances, it additionally beat the previous-generation RTX 3090.

I feel the candy spot for a card like this might be round $600-700, however with a MSRP of $799, it’s significantly cheaper than both the RTX 4080 or RTX 4090, making it straightforward to suggest to artists on a funds. Its important limitation is its 12 GB of graphics reminiscence, so when you do memory-intensive work, and might discover a secondhand GeForce RTX 3090 for $600-800, which may be a greater choice.

 
General conclusions
Whereas all three GeForce RTX 40 Collection playing cards on check are monsters, my very own choose could be the RTX 4090. It’s a improbable GPU for content material creation, particularly modelling and rendering, and its 24 GB body buffer is massive sufficient for all however essentially the most demanding DCC duties. Simply be sure you have a big sufficient PSU to maintain it fed: whereas 850W is the minimal, I’d suggest 1000W to depart room for overclocking and a quicker CPU.

My runner-up choose is the RTX 4070 Ti. Though it’s fairly a bit much less highly effective than the RTX 4090, it’s higher worth for cash than the the RTX 4080. Its 12 GB of GPU reminiscence will probably be a limitation in some workflows, however as software program builders get higher at optimising out-of-core efficiency, that will change into much less of a difficulty.

Lastly, I need to say thanks for taking the time to cease by. I hope this evaluate has been useful and informative, and when you’ve got any questions or solutions, please let me know within the feedback.

Hyperlinks

Learn extra in regards to the GeForce RTX 40 Collection GPUs on Nvidia’s web site

In regards to the reviewer

CG Channel {hardware} reviewer Jason Lewis is Lead Surroundings Artist at Staff Kaiju, a part of Tencent’s TiMi Studio Group. See extra of his work in his ArtStation gallery. Contact him at jason [at] cgchannel [dot] com

Acknowledgements

I want to give particular due to the next individuals for his or her help in bringing you this evaluate:

Stephenie Ngo of Nvidia
Sean Kilbride of Nvidia
Jordan Dodge of Nvidia
Chloe Larby of Grithaus Company
Lino Grandi of Otoy
Stephen G Wells
Adam Hernandez

Tags: 12VHPWR, 3DMark, 3ds max, Ada Lovelace, Adobe, Ampere, animation, arch viz, architectural visualization, Structure, Arnold, Arnold for Maya, Asus, Asus ProArt Studiobook 16, autodesk, Axiom, base clock velocity, benchmark, benchmark rating, benchmark check, finest RTX GPU, Blender, enhance clock velocity, CAD, cg, Chaos, Chaos Vantage, CineBench, Cinema 4D, clock velocity, compute energy, client card, client GPU, CryEngine, CUDA, present draw, Cycles, D5 Render, DCC, DirectX Raytracing, DXR, encoding, Enterprise Driver, Epic Video games, fluid simulation, kind issue, Founders Version, Foundry, FP32, FP64, Fusion 360, recreation artwork, recreation improvement, Sport Engine, gaming card, gaming GPU, gaseous fluid simulation, GDDR, GDDR6, GDDR6X, GeForce RTX 20 Collection, GeForce RTX 2080 Ti, GeForce RTX 30 collection, GeForce RTX 3070, GeForce RTX 3090, GeForce RTX 40 Collection, GeForce RTX 4070 Ti, GeForce RTX 4080, GeForce RTX 4090, GPU, GPU rendering, graphics card, graphics reminiscence, group check, hardware-accelerated ray tracing, Houdini, Jason Lewis, KeyShot, LuxCoreRender, Luxion, Maverick Studio, Maxon, Maya, reminiscence bandwidth, reminiscence sort, Metashape, modeling, modo, mtoa, NVIDIA, OctaneBench, OctaneRender, Omniverse Create, OptiX, Otoy, overclocking, P3 Kill A Watt Meter, energy consumption, Premiere Professional, worth, PSU, Pyro, Quadro RTX 8000, ray tracing, ray tracing core, Redshift, rendering, rendering benchmark, Assessment, RT core, RT cores, RTX, RTX A6000, SideFX, simulatino, simulation, Solidworks, Studio driver, Substance 3D Painter, artificial benchmark, TDP, Tensor core, Tensor cores, texturing, thermal design energy, Titan RTX, Turing, Unigine, Unity, Unreal Engine, V-Ray, V-Ray GPU, worth for cash, vfx, viewport benchark, viewport efficiency, digital actuality, visible results, visualize, VR, which GeForce RTX 40 Collection GPU is finest for CG work, which GeForce RTX 40 Collection GPU is finest for content material creation, which RTX GPU to purchase, workstation card, workstation GPU, Xidax