RTX 4060 vs ARC A770 vs RTX 2080 Super (Benchmark in 9 Games at 1440p) 2024

[@vickiepleiss5538]

Thank you for this comparison. I was checking another comparison video and couldn't figure out how the A770 was performing so much worse than the 4060, because I have used both, and the A770 consistently performed better than the 4060 on my system in many newer titles. Turns out the other comparison video had ReBAR off.

[@AndrewMorgan-c5h]

Great to see the Arc card trading blows with the 4060 but for me the srar of the show here was the 2080 super. I nean it just goes to show how good the card was at launch.

[@teflonstestbench]

Completely agree, in fact it might be the cheapest of these 3 too.

[@piyushthenub]

What's srar?

[@yokingboii]

@@piyushthenub he meant star

[@petera8665]

2080S uses twice as much energy as 4060 and A770 uses 50% more than 4060. No card beats 4060 on FPS/Watt. That is impressive!

[@AndrewMorgan-c5h]

@@petera8665 true, but one is 2 generations old and the other is a first generation product. The fact that at more money than either of them the 4060 isn't able to outperform those cards consistently is disappointing at best.

[@LapocilgadelEmi]

Thanks for the benchmark bro.💪

[@teflonstestbench]

Thanks for the support :)

[@0utkastmedia922]

great benchmark video but may i suggest doing it again but with the 4060ti 8gb as both the 2080s and a770 are more expensive then the normal 4060 but are roughly the same as the 4060ti 8gb

[@teflonstestbench]

Yup, I have a RTX 4060 ti right now and i will compare it to them in future videos :)

[@timschalau8425]

is it? bought my a770 bifrost for 300 bugs in sale

[@HaRvleee]

My arc a770 was cheaper then 4060 when I bought it

[@DmitriyPaliy]

The a770 memory frequency is not increased. I have the LE version and have the same 2187 MHz and 2400 core. But I changed the power limit on the 210 watt and made 2450 for the core. It works a bit better. Almost as 3070

[@teflonstestbench]

The memory frequency is increased, it is 2000mhz base, but half the models have a 2187 MHz one including the Intel LE and the Acer Bifrost, but there are models that do not as well like the Gunnir ones or the 8GB Intel one.

[@DmitriyPaliy]

@@teflonstestbench yes, because they have less price. Do you think the memory was overclocked by factory or is it normal datasheet frequency ?

[@teflonstestbench]

@DmitriyPaliy it's factory overclock, the datasheet one is 2000 MHz according to tech power up

[@TypeKraken]

Nice benchmark!

[@eajayy8969]

Honeslty appreciate the recency of this benchmark. Im thinking of putting my 2080ti in another rig (I guess my better rig) and switching back my a770 I put on hold for a couple months. Seems it performa at least 4060 performance

[@CalienW2]

0:41 i've seen a lot of videos/articles showin several performance lost on Arc when ReseizeBAR was disable. Is it real?

[@teflonstestbench]

I did my own video with the A750 and yeah, you do lose a lot of performance, that is why I always test with rebar on

[@sjtechzone]

I am planning to build a pc with Arc A770 for freelance video editing. I will mostly use Premiere Pro, After Effects and sometime Blender. So can you tell me which one will perform better in video editing gunnir or limited edition when LE have more bandwidth (560gbps vs 512gbps) and Gunnir have higher clock speed and better cooling system. I heard LE has heating problem. Is it solved now or does it still have a heating issue? If you tested those cards recently for video editing then let me know which one performs better in real-life video editing.

[@teflonstestbench]

Hey, the difference between those versions are really small, both have the oc memory. I personally had the LE version of the 770 and it was hot, not extremely hot but i needed to manually set the fans higher to keep it under 80c. I would probably go for the Gunnir one as if it is the 3fan version from them it seems like it will have better cooling. I didn't test those cards for productivity but for video editing on premiere I find the CPU to be a lot more important as the GPU only helps when encoding at the end and accelerating some specific effects that support GPU acceleration.

[@sjtechzone]

@@teflonstestbench Thanks for sharing your experience. Premiere Pro and After Effects rely heavily on GPU for real-time playback. Which is the most important thing in video editing and obviously, a good CPU will increase performance in CPU-accelerated tasks and increase performance in multitasking. But according to my knowledge, I think investing in powerful GPU will really increase the video editing performance if I equip it with a decent CPU like Ryzen 5600 or 5700x.

[@teflonstestbench]

As I said before video playback is processed by the CPU and supported effects by the GPU, but most of the time is only the CPU, at least in all the projects I've done. And that seems to be the case as other people comment, see here for example: www.reddit.com/r/premiere/comments/10zbmm2/comment/j83mlot/?

[@sjtechzone]

@@teflonstestbench Actually, I did a small research about it and able to understand a good GPU is crucial for complex video editing. Cause GPU process the main video footage, overlay image, text and videos and also GPU process lumetri color effects for color correction and grading and some effects and transition is also process by GPU. It also process the keyframe animation during playback. For CPU effects or transition CPU calculate it and send the data to GPU and GPU show the data in real life by help of CPU. So GPU is mainly responsible for showing edited frame in real time and having a powerful GPU is really helpful for complex editing.

[@budge1972]

Nice one, thankyou

[@NitpickZzz]

MSAA absolutely slaughters the arc cards for some reason

[@teflonstestbench]

Are you sure it's due to MSAA? I mean I haven't tried other options to see if that is but it would be weird for that to be the cause

[@NitpickZzz]

@@teflonstestbench Yes pretty much. It's a prevalent issue in GTA 5 as well where it can pull 200+ fps on maximum settings with Msaa disabled but gets about 50-60 with Msaa enabled

[@p_mcg]

Definitely MSAA ​@@teflonstestbench

[@teflonstestbench]

@p_mcg but still MSAA is only on Forza Horizon and it didn't do much better in other games.

[@TrumpWorshipper]

You get around twice as much performance per watt with the 4060! Impressive leap in power efficiency.

[@oleglevanzin6891]

Absolutely agree with you! But 8 gigabytes versus 16. This is a dead end, it has no future.

[@VoornePuttenBenchmarks]

Do you think the 4070 ti super is worth a 240€ increase over the 4070 super? Im currently unsure which one to get, the performance difference isn't that much at just 10 to 15% on avg but the ti super has 4gb more vram and a bigger memory bus which will result in a longer lifespan which is why im hesitant to buy the 4070 super.

[@teflonstestbench]

If you want to play at 4k them yes, if it is for 1440p then I think the 4070 super is good on price point and performance. 240€ difference I think it's fair between the two really, as despite the naming they have different chips and as you said, a bigger memory bus that will likely make it last longer. So as I said at the beginning I think it depends on the resolution you are targeting.

[@P1ayD0hpaul]

I’m building a budget 4k video editing rig, and read someone say the A770 performs very well for that. I don’t game at all. Would you recommend the A770 over its competitors in this video for rendering and encoding?

[@teflonstestbench]

The A770 has really good encoding capabilities, but I'm not sure about rendering. Rendering usually relies a lot more on the CPU and only accelerates some tasks on the GPU. I would look up for the cards that the program developers recommend, if they support ARC cards correctly on what you are using then I would go for that.

[@ph8808]

Intel gpus are a great choice for video encoding in general, I would go for it, maybe even a lower end model if you can get a good price on it, especially for AV1

[@GearGolemTMF]

How'd you show the real GPU power in the metrics? I've been wondering this for a while. I figure its more than what the normal metrics show.

[@teflonstestbench]

That power is calculated by an external device you plug on the PSU cables before the GPU, so it measures all the energy passing through them. I got mine from elmorlabs. The real power can vary and it might be more or less than the reported by the BIOS of the card depending on how each card calculates the value as some are approximations, other just calculates the power going to the core and memory and not the rest of the card...

[@GearGolemTMF]

@@teflonstestbench Thanks for the response! I used to wonder if I was seeing a lower number than the stock number due to not having enough power in my PSU. That makes sense.

[@teflonstestbench]

Well I don't think the PSU could limit the power in that way, in most of the cases the PSU will turn off if it went over the limit so if it is not getting more power it is because the GPU is not actually "requesting" more, so you should be fine. What is your GPU btw? And your psu?

[@east-e3s]

Hello. Thank you for the video. Can you perform benchmarks at 4K? To fully utilize the 16GB VRAM performance of the A770, running at 4K would be ideal. The A770 shows disappointing chip design and low performance, but it has very high potential in a 4K environment and offers great efficiency for the price. Additionally, there have been significant performance improvements due to ongoing graphics driver updates that cannot be ignored. For more efficient viewing, please use timestamps to divide the video into sections by topic. This is an editing feature used by most benchmark reviewers and many KZbinrs today. The specifications for benchmarking should be include information about the cooling system, monitor specifications, and Windows version. There is a significant performance difference between Windows 10 and 11. The specifications of the benchmark video do not include the Windows version or the Windows update version. I am curious whether the INTEL A770 is a reference or non-reference model, as there is a considerable performance difference between the INTEL reference version of the A770 and the ASROCK non-reference version. It is great that you included the updated version of the graphics card in the specifications section. However, using a non-latest graphics driver version for benchmarking is a critical mistake. The NVIDIA graphics driver version 556.12 and the INTEL graphics driver version .6078 are already outdated. I don't know when the video was made, but as of the upload date of September 30, 2024, both graphics driver versions are already one term late. Of course, it may not be a big issue, but depending on the update information, it can be a variable that significantly affects performance. For a more professional benchmark review, it is essential to measure and categorize variables in various situations. Most importantly, the specifications of the AMD R7 5800X, DDR4-3600MHz, and QuadHD (2560 x 1440) monitor are particularly unsuitable for benchmarking, and I find your benchmark review lacking in professionalism, which is quite disappointing. Of course, considering the hardware specifications of general users who may find high-end hardware burdensome, it might not be a bad review, and there may not be a need to evaluate amateur reviewers against high standards. However, for viewers seeking professional benchmarking information from reviewers or KZbinrs, it provides limited information and unprofessional analysis. But it was not a bad review. Thank you. From Korea.

[@teflonstestbench]

Hey, thanks for taking the time to write such a long and detailed comment with feedback, those are always welcome as I'm always looking to improve, so let me respond to it: Firstly, I don't think I will test the arc a770 at 4k, not because of the VRAM but because it does not have enough core performance. The card struggles even at 1440p ultra gaming in some titles, so jumping to 4k would be a slideshow not because of VRAM constrains but because of the core nor being powerful enough. Maybe for 4k low-medium, but that would use less VRAM than 1440p ultra due to texture quality. About the timespans I forgot to include them in this video and actually I think they are not on the latest video either, when I have some time I will log into my PC and do it, but thanks for the reminder, I try to add them to all the videos but sometimes forget. About the benchmarking information I try to add as much as possible in the intro of the video, like modifications I've made to the configurations of the card if I have done any and other observations. Also I try to give as much information about the testing environment and testbench but it is a good idea to add the cooler I use for the CPU too, so I will do that for the next vid on. About the windows version I include the main branch I'm on, but the minimal versions and updates are harder to list and find and I think not really that useful for most people so that is why I don't bother to do it. Also about the w10 vs 11 thing I prefer Windows 10 because it is a lot more stable than w11 when it comes to GPU and CPU performance as even now there are updates and features on there affecting performance, but I will move to w11 in the future when I get some money to upgrade the testbench to am5. Also what I have seen about this is that w11 is not much faster than w10 when benchmarking, differences coming down to some specific games or hardware configurations like the latest CPUs so didn't feel the need to change that fast. The GPU versions is something I list clearly at the beginning of the video, for the arc is the Acer Bifrost one as you can see in the first parts of the intro, after i list the games tested along with a screenshot of GPU-Z. About the drivers something special happened when making this video, it was scheduled for a month before i released it so that is why it has that driver version. I usually always use the latest WHQL driver available for each card. The reason it is not for this video is because of how I manage my cards: In order to test them I have to buy the GPUs with my own money, that means that because I'm a 22 year old student I don't have much money around, so when I want to buy a new GPU I'm forced to sell one I already have second hand. In this case I sold the RTX 2080 super to buy the arc a770, but the thing is that the arc I got had a problem and did not work, so I had to resolve the issue and that took 4 weeks until I got another, so that is because the driver versions are old for the Nvidia cards, as I always test the same brand with the same drivers to ensure all cards are tested in the same conditions in each video. And that is also why the arc had the latest at that moment. About my specifications I don't have enough money to afford the latest hardware to test on, as I'm not given anything by brands like most professional reviewers due nor I get any monetary benefit from this tests rather than KZbin ads that due to how taxes work on my country I haven't even gotten yet, so I know it's not the best but at least I ensure all cards are tested in the same situation as that is more than enough to show the differences between them in performance if the GPU us the only variable that changes. About the monitor I use a virtual one via a capture card, so I can configure up to 4k 30 if I need to, tho I don't think that I will test at 4k due to not having enough power to properly edit footage that big. Also I don't consider myself a professional reviewer, tho I would like to be one in the future. That doesn't mean I do a bad job, in fact with what I have I think I do a way better job than some professional ones that put no care on the tests they do or test GPUs with different GPUs or just make up results all together. That is why I'm all the tests I do I show the full footage of the game running so you can see exactly what is happening in real time and judge by yourself if the benchmarks are good, instead of doing what a lot of people do that is just writing down a number you don't know where they got it from. Again thank you so much for the feedback, I really appreciate it as I'm looking to improve and if you have any other suggestions about how I could improve or comunicate better I would love to hear them :)

[@Hetsu..]

@@teflonstestbench Man keep up the good work bro

[@kennethwilson8236]

Yeah the A770 will outperform the 4060 at 1440p and 4K

[@fblmagic6559]

Is the a770 god for 1080p max graphics?

[@teflonstestbench]

Yep, both ARC 750/770 are really good for those settings and will run most games well, even at 1440p tho there you can run into more problems depending on the optimization.

[@fblmagic6559]

@@teflonstestbench thanks for the fast reply. Just subbed 🐐

[@fblmagic6559]

@@teflonstestbench also is the 5600 cpu good to pair with the a770 or should i go all intel?

[@eltiopoder]

4060ti vs arc a770 16gb

[@teflonstestbench]

Will do next vid

[@hvdiv17]

Team blue lets gooooooo!!!!!

[@NotSure1985]

Arc $220 right now

[@псю]

here comes the power consumption bozos

[@warfarenotwarfair5655]

"Power consumption" is for poor people and foreigners. Americans are not concerned with a few measly watts especially under a Trump economy where the energy sector can breathe again.

[@aisuotome1413]

2080 Super = 3060 Ti 4060 = 2080

[@FagundesPinturas]

Parabéns para a intel, conseguiu criar um produto, que consome mais, tem menos desempenho, e custa mais caro

[@warfarenotwarfair5655]

Nobody in America cares about power consumption and quick sync is better than anything Nvidia can produce. Not everyone is just buying GPUs for 100% gaming 😂. The average GPU statistics on Steam are garbage integrated laptop GPUs.

[@NoSpamForYou]

It looks like intel designed ARC to do good in dedicated benchmarks and flubbed actual game performance.

[@teflonstestbench]

It is that, optimizing takes time and effort, so it makes sense to optimize for benchmarks. Game performance is hard to optimize as there are a lot of games and engines to optimize to, but with time they have been making improvements and increasing the game performance and stability in many titles.