DUDE, great stuff! In a world of sketchy benchmark channels you're not one of them. Your website is an awesome resource! The fact that you list the API your using in the tests AND give disclaimers of less known features being enable but not always fully supported like rebar/sam is a good indicator that you actually know what you're talking about. My only critique is to include the driver versions of the gpu and the game you are testing. Drivers can affect performance so much so it would be good to know. Subscribed!
Thank you so much for the nice comment and the feedback. The driver version I put it in the intro of the video and also in my website in the benchmarks page, adding it to the test footage itself I don't think it's necessary as it stays the same through the whole test batch and each GPU, that's why I put all that extra information at the beginning of the tests so if you are interested you can check it out. About the game version is something I was thinking about but it's a little bit hard as a lot of games don't show the version they have clearly or are constantly updating so it makes it hard to track them. What I can assure is that all the tests in the same video are in the same conditions, same game version and system configuration for all of them, and I take it really seriously and I had to scrap some footage in the past due to games changing and updating mid testing xD
@@teflonstestbench Oops, my b I missed that line on the drivers. I've done a couple benchmarks myself and there is so much to track so I appreciate all that info. More importantly it tells me you're legit. Keep up the good work mate!
Np, as you say there is a lot of info to show so it's actually usefully to know if you missed the drivers thing so I can try to place it better or change the format so other people can find it more easily, so thanks for the feedback really :) And yeah, benchmarks are easy to make but hard to make right as there are tons of info and especially understanding what it means and how it reflects on each product and scenario is hard, I was thinking about doing a video explaining it a bit as I understand that this kind of benchmark Is less attractive that what other people do of just putting the avg framerate number and calling it a day. But that is why I started doing benchmarks and uploading them, I wanted to know the performance but most places gave weird data or non verificable one so I though I may just as well do it and post it for others to use it too, and I would love to keep calling this and do a lot more tests with a lot more parts and also upgrade the website so it goes along with the channel and can be a database to check and compare all the data I post here
First-time viewer here. This was a really good benchmark! Thanks for putting in the hard work. I know it's best practice to show the cards physically to prove you have the hardware, but I'm pretty sure this is legit. If someone is buying today, the 7600 clearly seems to provide the best value for money.
Thank you so much, to be honest I didn't think about putting pics of the cards on the video, but I will and I think I did it some time ago in some benchmarks. The thing with that is that even if putting the pictures you can fake the results just putting other card instead, that's why I put all the metrics and data so you can check the VRAM usage, speeds, power consumption and pretty much everything and I'm always open to showing the raw footage were I always have GPU-z open to show the cards I've got. Tho now that I think about it a GPU-z screenshot maybe could be good for verifying.
Im confused what shoud i but an rtx 3070 or rx 6700xt i play games and use productivity software like after effects,photoshop,blender as hobby and side husstle,im leaning towards 3070 but i see the recent videos and it seems 8 gb vram aint gonna hold for long ( getting both card at same price).
For 1080p the RTX 3070 seems to be a little better and Nvidia is more productivity friendly, but for 1440p the VRAM will be a problem in the future for the RTX card so if you plan on gaming at that resolution then the RX 6700 XT is the better option
@@teflonstestbenchrx6700xt is available at 85k PKR and rtx3070 is at 90k, that's about $20 difference For 1080p gaming in the future for upcoming titles like doom dark ages, black myth wukong and Phantom Blade 0, which of the two should I pick Again I intend on 1080p gaming with NO ray tracing
Then RTX 3070 is the best fit for that, as the 8GB should be enough for 1080p and ray tracing for the years to come. I don't know if a specific game will need a little bit more but the performance loss because of that will be minimal anyways so the RTX 3070 is better, especially if the difference is only 20$.
@@teflonstestbench interesting, so brother, you mentioned "a specific game might need more" I assume you mean VRAM, in that case shouldn't I go for the 6700xt. Please don't be bothered by me asking as I'm expanding my knowledge here. Another channel recommend the 6700xt saying it has more VRAM and for 1080p gaming it'll be better for future games. *Now my CAVEMAN brain is confused again*
Yeah I was kinda expecting this question after I said that in my response. The thing about VRAM in games is that when you run out of it the performance doesn't just suddenly drop massively, but instead you start bottlenecking the GPU core a bit by making it use virtual RAM for the data it needs instead of the actual VRAM which is a lot faster. But the GPU (driver) has a memory management system so it will prioritize the thing it needs the most for the VRAM and cycle the other stuff that is important so it can get the best performance. That is why I'm taking some tests in my videos you see that the value next to the VRAM used is sometimes big even if it has VRAM free. (That value indicates the shared memory the GPU is using) So basically even if the VRAM fills up it will still perform good, it will lose a couple of potential frames it could be processing if it had bigger ram, but the game won't go from playable to unplayable unless the difference is really big. For example if the GPU has 8GB but is using 8.5GB it will be fine. If it needs to use 10GB there will be a problem and then depending on the game, engine or how the driver handles it then you can have big dips in performance. So summing up, 8GB should be enough and even if you suddenly need a little big more it will still be fine. The thing is that most games will be fine with that capacity and even if a buggy or unoptimized game releases that uses a little bit it will still be perfectly fine and probably have more performance still than the RX 6700 XT, but that situation is so so unlikely I would personally not account for it and I would go for the RTX 3070. I hope you have a better idea now and if you are still confused about it ask me anything you need. This explanation is a little bit simplified and not exactly like GPUs work but that is such a huge topic that it's hard to summarize it up on a comment, but I'll add it to my video ideas list as it would be really interesting to explore indepth.
Interesting thing I noticed was the GPUs power consumption aligned with the average fps you showed. Funny that the 6700 drew more wats and so had more FPS as the 7600 drew less and thus lower FPS. What if they both consumed the same amount and worked on equal consumption?
I think what you are pointing out has to do with the design of the core, that doesn't seem like it changed a lot from RDNA2 to RDNA3. I realized that same thing with the RX 7600 and Rx 6600XT in another video that have almost the same power consumption and performance compared to one another and also similar memory configurations and are both the entry level of each of their generations.
Guess I'll hang on to my 3070 :) Interesting to note that the performance of the three cards seems to scale directly with their power consumption in Watts.
@@teflonstestbench No kidding, two opposing regimes reached similar conclusions when converting watts into frames kinda like the Allies and the Krauts developing jet engines in the 1940s :)
amazing video, i am planning to get either 4060ti 8gb or rx 6750xt which do you think is better? i am concerned that the rx 6750 might have stutters issues does this happen?
Thank you so much for the nice comment. About the Stutters they should not happen as this card is powerful enough and it has enough VRAM for 1080p and 1440p gaming. I have been using the RX 6700 XT for a while now that is almost the same and I don't remember seeing any stutters in any game I tested on it, so nothing to worry about. Stutters usually happen when bottlenecked by VRAM or RAM or when the CPU gets full. When it happens because of the GPU it will be because drivers or a buggy game and that can happen with any GPU and any game and is not exclusive to any GPU vendor or model, tho there were some that were more problematic than others it's impossible to predict or say when or in what scenarios will happen with them. So basically summing up both options are equal on that regard. If I must recommend one I think the RTX 4060 ti 8GB is better than the RX 6750 XT in terms of raw performance and it's more modern, the only thing is that VRAM is a problem with only 8GB its only suitable for 1080p gaming, and if you go for the 16GB version you will have a terrible VRAM bottleneck due to the 128 bit bus. The RX 6750 XT is much more balanced on that front, with 12GB of VRAM it will be perfect for 1440p and 1080p, the only problem is that rat tracing performance is far from Nvidia's latest gen but if you are looking at just gaming and not that much into playing a really RT intensive game with RT at max I would go for the AMD option from those two.
I'm gonna buy RX 6700XT to ugprade my gaming rig from 1050ti, it only cost me 249$ Include changes the PSU to 650W Bronze 80+ Using PK650D Deep Cool Variant, :))))
Yeah but keep in mind that Nvidia and AMD use VRAM differently on their GPUs, you can notice that by looking at the rest of the games. And in that test the RTX 3070 doesn't seem to be using all the VRAM available still and is not stuttering or having any problem, so maybe on other games it will have problems but for RDR2 in particular it seems to be good.
I don't think testing PUBG is a good idea as it is really CPU limited specially on that scenario so all cards would get similar results + it would be hard to get usable results due to each drop being different and loading/rendering different things
It is turned on on every compatible card, both ReBAR for Nvidia and SAM for AMD. Tho despite the differences in naming they are equivalent technologies and do the same thing.
![Tee Grizzley - Detroit (Feat. 42 Dugg) [Official Video]](http://i.ytimg.com/vi/aeKfTtt0x14/mqdefault.jpg)
DUDE, great stuff! In a world of sketchy benchmark channels you're not one of them. Your website is an awesome resource! The fact that you list the API your using in the tests AND give disclaimers of less known features being enable but not always fully supported like rebar/sam is a good indicator that you actually know what you're talking about. My only critique is to include the driver versions of the gpu and the game you are testing. Drivers can affect performance so much so it would be good to know. Subscribed!
Thank you so much for the nice comment and the feedback. The driver version I put it in the intro of the video and also in my website in the benchmarks page, adding it to the test footage itself I don't think it's necessary as it stays the same through the whole test batch and each GPU, that's why I put all that extra information at the beginning of the tests so if you are interested you can check it out. About the game version is something I was thinking about but it's a little bit hard as a lot of games don't show the version they have clearly or are constantly updating so it makes it hard to track them. What I can assure is that all the tests in the same video are in the same conditions, same game version and system configuration for all of them, and I take it really seriously and I had to scrap some footage in the past due to games changing and updating mid testing xD
@@teflonstestbench Oops, my b I missed that line on the drivers. I've done a couple benchmarks myself and there is so much to track so I appreciate all that info. More importantly it tells me you're legit. Keep up the good work mate!
Np, as you say there is a lot of info to show so it's actually usefully to know if you missed the drivers thing so I can try to place it better or change the format so other people can find it more easily, so thanks for the feedback really :)
And yeah, benchmarks are easy to make but hard to make right as there are tons of info and especially understanding what it means and how it reflects on each product and scenario is hard, I was thinking about doing a video explaining it a bit as I understand that this kind of benchmark Is less attractive that what other people do of just putting the avg framerate number and calling it a day. But that is why I started doing benchmarks and uploading them, I wanted to know the performance but most places gave weird data or non verificable one so I though I may just as well do it and post it for others to use it too, and I would love to keep calling this and do a lot more tests with a lot more parts and also upgrade the website so it goes along with the channel and can be a database to check and compare all the data I post here
bought a 6700xt for $325CAD and its arriving today! been using a 1070 since 2018 and cant wait for the huge increase
hey its been 3 months, how has the 6700 xt serving you so far?
First-time viewer here. This was a really good benchmark! Thanks for putting in the hard work. I know it's best practice to show the cards physically to prove you have the hardware, but I'm pretty sure this is legit. If someone is buying today, the 7600 clearly seems to provide the best value for money.
Thank you so much, to be honest I didn't think about putting pics of the cards on the video, but I will and I think I did it some time ago in some benchmarks. The thing with that is that even if putting the pictures you can fake the results just putting other card instead, that's why I put all the metrics and data so you can check the VRAM usage, speeds, power consumption and pretty much everything and I'm always open to showing the raw footage were I always have GPU-z open to show the cards I've got. Tho now that I think about it a GPU-z screenshot maybe could be good for verifying.
Picked up an open box 7600 for $225 for my 2nd pc...best option for $200 range card for now.
Upgraded from an RX 6600 to an 7600 and the difference is quite noteworthy! Only costed me a 180€ too, got quite lucky i guess!
How much different?
Im confused what shoud i but an rtx 3070 or rx 6700xt i play games and use productivity software like after effects,photoshop,blender as hobby and side husstle,im leaning towards 3070 but i see the recent videos and it seems 8 gb vram aint gonna hold for long ( getting both card at same price).
For 1080p the RTX 3070 seems to be a little better and Nvidia is more productivity friendly, but for 1440p the VRAM will be a problem in the future for the RTX card so if you plan on gaming at that resolution then the RX 6700 XT is the better option
@@teflonstestbenchrx6700xt is available at 85k PKR and rtx3070 is at 90k, that's about $20 difference
For 1080p gaming in the future for upcoming titles like doom dark ages, black myth wukong and Phantom Blade 0, which of the two should I pick
Again I intend on 1080p gaming with NO ray tracing
Then RTX 3070 is the best fit for that, as the 8GB should be enough for 1080p and ray tracing for the years to come. I don't know if a specific game will need a little bit more but the performance loss because of that will be minimal anyways so the RTX 3070 is better, especially if the difference is only 20$.
@@teflonstestbench interesting, so brother, you mentioned "a specific game might need more" I assume you mean VRAM, in that case shouldn't I go for the 6700xt. Please don't be bothered by me asking as I'm expanding my knowledge here.
Another channel recommend the 6700xt saying it has more VRAM and for 1080p gaming it'll be better for future games.
*Now my CAVEMAN brain is confused again*
Yeah I was kinda expecting this question after I said that in my response. The thing about VRAM in games is that when you run out of it the performance doesn't just suddenly drop massively, but instead you start bottlenecking the GPU core a bit by making it use virtual RAM for the data it needs instead of the actual VRAM which is a lot faster. But the GPU (driver) has a memory management system so it will prioritize the thing it needs the most for the VRAM and cycle the other stuff that is important so it can get the best performance. That is why I'm taking some tests in my videos you see that the value next to the VRAM used is sometimes big even if it has VRAM free. (That value indicates the shared memory the GPU is using)
So basically even if the VRAM fills up it will still perform good, it will lose a couple of potential frames it could be processing if it had bigger ram, but the game won't go from playable to unplayable unless the difference is really big.
For example if the GPU has 8GB but is using 8.5GB it will be fine. If it needs to use 10GB there will be a problem and then depending on the game, engine or how the driver handles it then you can have big dips in performance.
So summing up, 8GB should be enough and even if you suddenly need a little big more it will still be fine. The thing is that most games will be fine with that capacity and even if a buggy or unoptimized game releases that uses a little bit it will still be perfectly fine and probably have more performance still than the RX 6700 XT, but that situation is so so unlikely I would personally not account for it and I would go for the RTX 3070.
I hope you have a better idea now and if you are still confused about it ask me anything you need. This explanation is a little bit simplified and not exactly like GPUs work but that is such a huge topic that it's hard to summarize it up on a comment, but I'll add it to my video ideas list as it would be really interesting to explore indepth.
Interesting thing I noticed was the GPUs power consumption aligned with the average fps you showed.
Funny that the 6700 drew more wats and so had more FPS as the 7600 drew less and thus lower FPS.
What if they both consumed the same amount and worked on equal consumption?
I think what you are pointing out has to do with the design of the core, that doesn't seem like it changed a lot from RDNA2 to RDNA3.
I realized that same thing with the RX 7600 and Rx 6600XT in another video that have almost the same power consumption and performance compared to one another and also similar memory configurations and are both the entry level of each of their generations.
Guess I'll hang on to my 3070 :) Interesting to note that the performance of the three cards seems to scale directly with their power consumption in Watts.
Yup, wasn't expecting the RX 6700 XT to be that close to the RTX 3070 at almost the same power consumption
@@teflonstestbench No kidding, two opposing regimes reached similar conclusions when converting watts into frames kinda like the Allies and the Krauts developing jet engines in the 1940s :)
amazing video, i am planning to get either 4060ti 8gb or rx 6750xt which do you think is better? i am concerned that the rx 6750 might have stutters issues does this happen?
Thank you so much for the nice comment. About the Stutters they should not happen as this card is powerful enough and it has enough VRAM for 1080p and 1440p gaming. I have been using the RX 6700 XT for a while now that is almost the same and I don't remember seeing any stutters in any game I tested on it, so nothing to worry about. Stutters usually happen when bottlenecked by VRAM or RAM or when the CPU gets full. When it happens because of the GPU it will be because drivers or a buggy game and that can happen with any GPU and any game and is not exclusive to any GPU vendor or model, tho there were some that were more problematic than others it's impossible to predict or say when or in what scenarios will happen with them.
So basically summing up both options are equal on that regard. If I must recommend one I think the RTX 4060 ti 8GB is better than the RX 6750 XT in terms of raw performance and it's more modern, the only thing is that VRAM is a problem with only 8GB its only suitable for 1080p gaming, and if you go for the 16GB version you will have a terrible VRAM bottleneck due to the 128 bit bus. The RX 6750 XT is much more balanced on that front, with 12GB of VRAM it will be perfect for 1440p and 1080p, the only problem is that rat tracing performance is far from Nvidia's latest gen but if you are looking at just gaming and not that much into playing a really RT intensive game with RT at max I would go for the AMD option from those two.
I have a 6750 XT, it's an awesome card
Hi Ive found 3070 and 6700 xt used at about $300 which one do you think is worth buying for me? upgrading from 1660 super
The RTX 3070 is faster so I would go for that, it should be fine for 1080p and 1440p if you don't use any really VRAM intensive games
@@teflonstestbench thank you so much!
I'm gonna buy RX 6700XT to ugprade my gaming rig from 1050ti, it only cost me 249$ Include changes the PSU to 650W Bronze 80+ Using PK650D Deep Cool Variant, :))))
Notice the GPU memory on RDR2 with the 6700XT.. Over 10GB of VRAM usage. The 3070 isn't keeping up despite higher frames
Yeah but keep in mind that Nvidia and AMD use VRAM differently on their GPUs, you can notice that by looking at the rest of the games. And in that test the RTX 3070 doesn't seem to be using all the VRAM available still and is not stuttering or having any problem, so maybe on other games it will have problems but for RDR2 in particular it seems to be good.
I wanted to see how AMD has come with their drivers
Firsr lol, btw i like 6700xt
can u please test pubg pc and do a hotdrop?
I don't think testing PUBG is a good idea as it is really CPU limited specially on that scenario so all cards would get similar results + it would be hard to get usable results due to each drop being different and loading/rendering different things
Tira as muleta deixa a placa trabalhar
3070 vs 7600xt please
I'll try to get the 7600 XT, but the only difference is the VRAM so the gains should be really small and they are hard to find for me locally
I suppose on amd cards Smart Access Memory was turned off and therefore low fps ?
It is turned on on every compatible card, both ReBAR for Nvidia and SAM for AMD. Tho despite the differences in naming they are equivalent technologies and do the same thing.