ASUS' Weird ROG Ally Non-Extreme: Z1 Benchmarks vs. Z1 Extreme Ally
video description
Date: 2023-10-10
Related videos
Comments and reviews: 20
aggies11
With respect to the GPU busy tool/metrics, I still think a better word than balanced might be useful. While the concept makes sense on it's own. In terms of actually what's happening, it's not balanced at all. You want as much of the frametime to be limited by the GPU, not CPU. That's really not balancing your components, it's making sure all the weight is on the GPU. The scale tips all the way to once side.
And it also doesn't tell us about the other direction. a Mid tier CPU that gives you 99% GPU busy vs a high end CPU that also gives you 99% GPU busy lead to the same metrics. Those are wildly different systems and so can't be the same balance of components, yet GPU busy would call them both balanced systems. It really does seem to be a metric that tells you how much the rest of the system is holding back the GPU, but not much about how much it isn't if that makes sense.
Still super useful info and a great metric, just might be worthwhile finding a better way of phrasing/describing/presenting it.
reply
With respect to the GPU busy tool/metrics, I still think a better word than balanced might be useful. While the concept makes sense on it's own. In terms of actually what's happening, it's not balanced at all. You want as much of the frametime to be limited by the GPU, not CPU. That's really not balancing your components, it's making sure all the weight is on the GPU. The scale tips all the way to once side.
And it also doesn't tell us about the other direction. a Mid tier CPU that gives you 99% GPU busy vs a high end CPU that also gives you 99% GPU busy lead to the same metrics. Those are wildly different systems and so can't be the same balance of components, yet GPU busy would call them both balanced systems. It really does seem to be a metric that tells you how much the rest of the system is holding back the GPU, but not much about how much it isn't if that makes sense.
Still super useful info and a great metric, just might be worthwhile finding a better way of phrasing/describing/presenting it.
reply
Sadnehs
I worry that the future of mini PCs and handhelds are going to reveal manufacturers cutting costs and corners in memory/cache bandwidth. Because it doesn't read easily on a sheet and doesn't hurt average fps.
But these devices will suffer stuttering and other more obtuse issues reviewers won't pick up on.
I would rather a device drop the overall performance in CPU / GPU, and invest in more stable performance in bandwidth. Yes I'd like to hit 60 fps, but if I can't I'd rather have a locked framerate.
Up in 144Hz territory I don't mind 10 or 20 fps swings. But down in the 40's - 30's, I do mind 5 fps, or a frame lasting 50% longer.
reply
I worry that the future of mini PCs and handhelds are going to reveal manufacturers cutting costs and corners in memory/cache bandwidth. Because it doesn't read easily on a sheet and doesn't hurt average fps.
But these devices will suffer stuttering and other more obtuse issues reviewers won't pick up on.
I would rather a device drop the overall performance in CPU / GPU, and invest in more stable performance in bandwidth. Yes I'd like to hit 60 fps, but if I can't I'd rather have a locked framerate.
Up in 144Hz territory I don't mind 10 or 20 fps swings. But down in the 40's - 30's, I do mind 5 fps, or a frame lasting 50% longer.
reply
Samira
In my experience there's something weird going on with Baldur's Gate 3's Vulkan implementation. There's weird hitching way too often and with ReBAR on the memory footprint just keeps going up (tried that spinning in place, no new assets loaded, footprint went up by 30GB at which point I quit the game, whereas on DX11 it reaches a point and actually decreases. Disabling ReBAR fixes the problem, except then you won't have it, and since DX11 works well in BG3... It's not like it's a hardware issue either, I'm running on a 5900X, RX 6800, 64GB DDR4-3200 and BG3 is installed on a fast PCIe 4 drive.
reply
In my experience there's something weird going on with Baldur's Gate 3's Vulkan implementation. There's weird hitching way too often and with ReBAR on the memory footprint just keeps going up (tried that spinning in place, no new assets loaded, footprint went up by 30GB at which point I quit the game, whereas on DX11 it reaches a point and actually decreases. Disabling ReBAR fixes the problem, except then you won't have it, and since DX11 works well in BG3... It's not like it's a hardware issue either, I'm running on a 5900X, RX 6800, 64GB DDR4-3200 and BG3 is installed on a fast PCIe 4 drive.
reply
Cqef
It sounds to me like the Ally was overpriced in the first place, then. The price difference between the Extreme and non-Extreme would have made more sense if they were priced even 100 bucks cheaper (600 for the Extreme, 500 for the non-Extreme). You'd be much closer to the Steam Deck in not only pricing, and the slightly higher pricetag of the Ally is more realistically representative of the performance bump you'd get with the Extreme, compared to how it is now (surely 700 isn't representative of how much better the Ally Extreme is to the decked-out Steam Deck, right?)
reply
It sounds to me like the Ally was overpriced in the first place, then. The price difference between the Extreme and non-Extreme would have made more sense if they were priced even 100 bucks cheaper (600 for the Extreme, 500 for the non-Extreme). You'd be much closer to the Steam Deck in not only pricing, and the slightly higher pricetag of the Ally is more realistically representative of the performance bump you'd get with the Extreme, compared to how it is now (surely 700 isn't representative of how much better the Ally Extreme is to the decked-out Steam Deck, right?)
reply
Watchman
Seeing these 1080p benchmarks, Nintendo ought to strive for 2k resolution at 60 frame rates per second, if capable; again, Nintendo should primarily focus on gameplay and perfect the 2k frame rates for a portable then focus on 4k. I want to play my games with frame rates and local coop play with friends than be concerned about irrelevant 4k hype. Perfect the gameplay and mid peak frame rate resolution and you got a primary gamer within me regarding portability and Nintendos vision
reply
Seeing these 1080p benchmarks, Nintendo ought to strive for 2k resolution at 60 frame rates per second, if capable; again, Nintendo should primarily focus on gameplay and perfect the 2k frame rates for a portable then focus on 4k. I want to play my games with frame rates and local coop play with friends than be concerned about irrelevant 4k hype. Perfect the gameplay and mid peak frame rate resolution and you got a primary gamer within me regarding portability and Nintendos vision
reply
Leap
I was honestly expecting a bigger gap in performance based off the specs. But unfortunately if you're playing 3D games on this thing, that extra 20% or so of performance can easily mean the difference between a game being fine, or flat-out unplayable. I can kinda see it being fine if you're mainly planning to play more lightweight/2D games on it though, which is probably the only case where you'll be able to take advantage of its upgraded screen anyways.
reply
I was honestly expecting a bigger gap in performance based off the specs. But unfortunately if you're playing 3D games on this thing, that extra 20% or so of performance can easily mean the difference between a game being fine, or flat-out unplayable. I can kinda see it being fine if you're mainly planning to play more lightweight/2D games on it though, which is probably the only case where you'll be able to take advantage of its upgraded screen anyways.
reply
moo
I don't really get how lower spec parts have lower boosts nowadays. It used to be that more cores meant lower frequency. Are these parts artificially downclocked so that higher / more expensive parts not only have more cores but also higher frequencies?
Btw, I held the Ally in my hands multiple times and every time it's plastic shell felt cheap compared to the GPD Win 4. I would even say that it felt closer to a 3D printed shell.
reply
I don't really get how lower spec parts have lower boosts nowadays. It used to be that more cores meant lower frequency. Are these parts artificially downclocked so that higher / more expensive parts not only have more cores but also higher frequencies?
Btw, I held the Ally in my hands multiple times and every time it's plastic shell felt cheap compared to the GPD Win 4. I would even say that it felt closer to a 3D printed shell.
reply
Delta
Wow, nice one ASUS. facepalm
I'm just curious why AMD cut the CN's by 2/3 from 12 to 4 rather than 1/2 from 12 to 6. Seems like a big hack considering cores only dropped from 8 to 6. Given that is has the same TDP, I would expect at least some clock speed is given back, but it isn't: it's reduced by 200 MHz. :/ All this is reflected in TFlops as listed on the spec sheet.
reply
Wow, nice one ASUS. facepalm
I'm just curious why AMD cut the CN's by 2/3 from 12 to 4 rather than 1/2 from 12 to 6. Seems like a big hack considering cores only dropped from 8 to 6. Given that is has the same TDP, I would expect at least some clock speed is given back, but it isn't: it's reduced by 200 MHz. :/ All this is reflected in TFlops as listed on the spec sheet.
reply
skitx31
Honest question here. Did you change the amount of VRAM the GPU can use? I know at default it is at 4GB but new firmware lets you up this to 6GB. If you did not I highly recommend you try it. I find this to be a good sweet spot for many games on my Ally. Yes I know there is a 8GB option as well but I kept running into lack of normal memory errors on a few games.
reply
Honest question here. Did you change the amount of VRAM the GPU can use? I know at default it is at 4GB but new firmware lets you up this to 6GB. If you did not I highly recommend you try it. I find this to be a good sweet spot for many games on my Ally. Yes I know there is a 8GB option as well but I kept running into lack of normal memory errors on a few games.
reply
Nine
probably not something you guys are interested in doing but i wonder how good a steam deck would be in combination with portable modems for live streaming on the go instead of buying something like a liveU. these liveU devices require you to have a server/computer somewhere else but with a steam deck this is no longer needed right? kinda gimmicky but still cool.
reply
probably not something you guys are interested in doing but i wonder how good a steam deck would be in combination with portable modems for live streaming on the go instead of buying something like a liveU. these liveU devices require you to have a server/computer somewhere else but with a steam deck this is no longer needed right? kinda gimmicky but still cool.
reply
Jeffrey
4 vs 12 CUs... based on the BENCHMARKS the MOBILE usage would seem to suggest that there's little reason to use more than SIX CU's?
A lot of games only got around 40% or so uplift with the 12 CU's vs 4 CU's. Having SIX CU's is 50% more than four so I'm not seeing much reason to go beyond six in general. Seems like the optimal balance there.
reply
4 vs 12 CUs... based on the BENCHMARKS the MOBILE usage would seem to suggest that there's little reason to use more than SIX CU's?
A lot of games only got around 40% or so uplift with the 12 CU's vs 4 CU's. Having SIX CU's is 50% more than four so I'm not seeing much reason to go beyond six in general. Seems like the optimal balance there.
reply
Tehpanda
there might be an interesting use case in super light workloads like handheld console emulation for the cut down chip. I am reasonably sure the battery life gain will be less than desired with the display choice, and an android device being suitable for this use case, but at least you get some more feature diversity and windows on the ally.
reply
there might be an interesting use case in super light workloads like handheld console emulation for the cut down chip. I am reasonably sure the battery life gain will be less than desired with the display choice, and an android device being suitable for this use case, but at least you get some more feature diversity and windows on the ally.
reply
itech
Id love to get the more powerful one. Super awesome to be abke to play nearly any game at playable levels.
A couple years ago i could have easily bought one... but with a family and everything costing so god damn much.... theres no extra money.
As long as thats my only problem... im ok. I hope everyone else is doing ok these days.
reply
Id love to get the more powerful one. Super awesome to be abke to play nearly any game at playable levels.
A couple years ago i could have easily bought one... but with a family and everything costing so god damn much.... theres no extra money.
As long as thats my only problem... im ok. I hope everyone else is doing ok these days.
reply
Daniel
I know previous generation apus saw some of the biggest gains from increasing interconnect speeds. If possible, it would be interesting to see the effects it has on this generation.
Edit: looked into this and there already is a tool called smokeless which should unlock the necessary options (fclk) to further tinker with these devices.
reply
I know previous generation apus saw some of the biggest gains from increasing interconnect speeds. If possible, it would be interesting to see the effects it has on this generation.
Edit: looked into this and there already is a tool called smokeless which should unlock the necessary options (fclk) to further tinker with these devices.
reply
Tomo
I've been waiting forever for Steve to show a some handheld reviews since the last Rog Ally video from several weeks ago since he said many more new handheld videos were coming but he instead gives another obvious Rog Ally video! Am I the only one not going to waste my time here watching what we already know? So disappointing.
reply
I've been waiting forever for Steve to show a some handheld reviews since the last Rog Ally video from several weeks ago since he said many more new handheld videos were coming but he instead gives another obvious Rog Ally video! Am I the only one not going to waste my time here watching what we already know? So disappointing.
reply
Francis
I think 4CU RDNA graphics could be reasonable inside a small, light, passively cooled case with the sole purpose of playing the majority of stuff in the GoG library. Basically everything up and until roughly the Windows 7 era? Would also be interesting to see what classics work fine 60fps on battery/power.
reply
I think 4CU RDNA graphics could be reasonable inside a small, light, passively cooled case with the sole purpose of playing the majority of stuff in the GoG library. Basically everything up and until roughly the Windows 7 era? Would also be interesting to see what classics work fine 60fps on battery/power.
reply
BBtech0251
The reason for much poorer performance for only 100 less is probably related to the naming confusion you mentioned at the beginning of the video: getting well-intentioned but uniformed parents to buy the wrong item at Christmas for a greater profit margin.
Also, did you get a new audio guy?
reply
The reason for much poorer performance for only 100 less is probably related to the naming confusion you mentioned at the beginning of the video: getting well-intentioned but uniformed parents to buy the wrong item at Christmas for a greater profit margin.
Also, did you get a new audio guy?
reply
TopPercent
Not asking for a video to address it, but other people who've used these handhelds have observed that downloading drivers from AMD tend to be more up to date and have better performance than the ones from Armory Crate. Have you had similar experiences or have you just used Armory Crate?
reply
Not asking for a video to address it, but other people who've used these handhelds have observed that downloading drivers from AMD tend to be more up to date and have better performance than the ones from Armory Crate. Have you had similar experiences or have you just used Armory Crate?
reply
watercannons
I always felt it was weird how large the iGPU gap was between the Ryzen 5 and Ryzen 7s for the 6000-series compared to the 5000-series. The 7000-Phoenix CPUs isn t as bad, but considering these generations are very much about the iGPU performance, it s a bit of a shame
reply
I always felt it was weird how large the iGPU gap was between the Ryzen 5 and Ryzen 7s for the 6000-series compared to the 5000-series. The 7000-Phoenix CPUs isn t as bad, but considering these generations are very much about the iGPU performance, it s a bit of a shame
reply
UngoKast
For the price, I couldn t imagine going with either version. Might as well go with something cheaper if you re just playing 2D games, or might as well go with something more powerful if you re trying to get above 60 fps on low settings, like a 13-15 inch gaming notebook.
reply
For the price, I couldn t imagine going with either version. Might as well go with something cheaper if you re just playing 2D games, or might as well go with something more powerful if you re trying to get above 60 fps on low settings, like a 13-15 inch gaming notebook.
reply
Add a review, comment
Other channel videos
