hahahaha
Even PS2 ports perform questionably, there will never be a playable PS2 emulator on Vita.
hahahaha
Even PS2 ports perform questionably, there will never be a playable PS2 emulator on Vita.
Assuming that there are no physical or practical hardware restrictions, we will probably have to wait for a kernel hack or perhaps even a lower-level firmware hack unless there is a setting available in user mode that would allow us to change the clock speed to unsupported frequencies. However, with that being said, there are a few specific statements that I would like to make concerning the need, usefulness, and practicality of an overclock.
First of all, the OP made a statement about the 2 ghz theoretical clock speed, and I too came across this number while looking at the hardware specs for the Vita on Wikipedia and several tech outlets. This 2 ghz clock speed was actually reported by ARM when they were describing a specific single/dual core 28 nm Cortex A9 manufactured by Global Foundries. I believe that it is most likely that the Vita will never be able to reach anywhere close to this speed because the SOC within the Vita was most likely not manufactured using the same process. This 2 ghz claim is most likely a similar hyperbole to when it was claimed by similar tech outlets that the Wii U's architecture was based on POWER7 because they were both multicore processors and they both had eDRAM (it later turned out that the Wii U was just running a modified tri-core PPC750 with the mere addition of eDRAM). Another item to take into consideration is that the Vita was most likely down-clocked not only to maintain a balance between energy and power consumption but also to be able to deliver sustained performance. Many smartphones on the market that advertise a highly clocked processor actually operate in a dynamic range based on the amount and type of workload fed to the processor as well as the heat generated. Thus, these smartphones often do not continually operate at the advertised frequencies. One example I will provide of sustained performance in smartphone SOCs is in the Snapdragon 810 used in the Nexus 6P. Anandtech found that the Nexus 6P was only able to operate at its maximum frequency of around 1.9 GHZ for 45 seconds after which it started to operate at around 1.9-1.8 GHZ, after 5 minutes it stabilized at around 1.3 ghz, and again after 12 minutes it dropped to 960 MHz, and finally the Cortex A57 core shutoff and the cooler Cortex A53 took over. This type of running a CPU as fast as possible but letting it throttle if under sustained pressure over a period of time makes sense in a smartphone, because a smartphone is generally designed to do a task as quickly as possible so that it can return to idle mode and conserve battery. However, this type of operation would be unacceptable in a portable system where predictability and sustained performance is desired by both the user and the developer (many people would not like the frame rate of their game to suddenly drop after a period of usage).
Secondly, I believe there are things beyond the clock speed within the CPU that can make a bigger difference. For example, IPC performance is important, as is how the microarchitecture enables instruction-level parallelism, what is available to process floating-point and SIMD operations, what speed the cache operating at, etc. All of this is also why I get a bit peeved when people directly compare something like the New 3DS's or the PSP's CPU with that of the PS Vita and claim the PSP offers the same performance or the New 3DS is faster simply because of the clock speed (I do not know if it is or is not faster, but you cannot just generalize about performance simply based on clock speeds). This last sentence does not have much to do with the topic, but I just needed to get that off of my chest after reading some very flawed arguments on this topic. However, my point is that there is so much more in the CPU and indeed within the overall system that affects performance. Depending on the microarchitecture of the CPU, overclocking (assuming it is possible) might even be detrimental to performance in certain cases; for example, maybe increasing the CPU speed beyond a certain point will result in one or more of the cores shutting off (which would be bad if the software you are running is optimized for multicore scenarios, but might be good if your software only uses one core).
Lastly, when it comes to emulators or apps/games in general, obviously beefier hardware helps and is even sometimes a necessity depending on what it being emulated or rum. However, with that being said, optimizations and programming skill can go a long way towards creating applications that perform well. In addition, it must be said that the Vita, like other dedicated consoles, offers developers a non-fractured base (everyone with a Vita gets a system with the same microarchitecture) and deeper hardware access, which leaves plenty of room for optimization. Furthermore, many of the emulators that are running natively on the Vita are poorly optimized and so there is a lot of room for overall improvement. As the OP said maybe a slight clock speed bump will make a small difference on some games and emulators (I very much doubt a huge overclock would be achievable or practically desirable even if it was possible), but I feel that at this point the focus should be on optimization. After all look at what was made possible on the much less capable PSP through optimizations.
First of all, the OP made a statement about the 2 ghz theoretical clock speed, and I too came across this number while looking at the hardware specs for the Vita on Wikipedia and several tech outlets. This 2 ghz clock speed was actually reported by ARM when they were describing a specific single/dual core 28 nm Cortex A9 manufactured by Global Foundries. I believe that it is most likely that the Vita will never be able to reach anywhere close to this speed because the SOC within the Vita was most likely not manufactured using the same process. This 2 ghz claim is most likely a similar hyperbole to when it was claimed by similar tech outlets that the Wii U's architecture was based on POWER7 because they were both multicore processors and they both had eDRAM (it later turned out that the Wii U was just running a modified tri-core PPC750 with the mere addition of eDRAM). Another item to take into consideration is that the Vita was most likely down-clocked not only to maintain a balance between energy and power consumption but also to be able to deliver sustained performance. Many smartphones on the market that advertise a highly clocked processor actually operate in a dynamic range based on the amount and type of workload fed to the processor as well as the heat generated. Thus, these smartphones often do not continually operate at the advertised frequencies. One example I will provide of sustained performance in smartphone SOCs is in the Snapdragon 810 used in the Nexus 6P. Anandtech found that the Nexus 6P was only able to operate at its maximum frequency of around 1.9 GHZ for 45 seconds after which it started to operate at around 1.9-1.8 GHZ, after 5 minutes it stabilized at around 1.3 ghz, and again after 12 minutes it dropped to 960 MHz, and finally the Cortex A57 core shutoff and the cooler Cortex A53 took over. This type of running a CPU as fast as possible but letting it throttle if under sustained pressure over a period of time makes sense in a smartphone, because a smartphone is generally designed to do a task as quickly as possible so that it can return to idle mode and conserve battery. However, this type of operation would be unacceptable in a portable system where predictability and sustained performance is desired by both the user and the developer (many people would not like the frame rate of their game to suddenly drop after a period of usage).
Secondly, I believe there are things beyond the clock speed within the CPU that can make a bigger difference. For example, IPC performance is important, as is how the microarchitecture enables instruction-level parallelism, what is available to process floating-point and SIMD operations, what speed the cache operating at, etc. All of this is also why I get a bit peeved when people directly compare something like the New 3DS's or the PSP's CPU with that of the PS Vita and claim the PSP offers the same performance or the New 3DS is faster simply because of the clock speed (I do not know if it is or is not faster, but you cannot just generalize about performance simply based on clock speeds). This last sentence does not have much to do with the topic, but I just needed to get that off of my chest after reading some very flawed arguments on this topic. However, my point is that there is so much more in the CPU and indeed within the overall system that affects performance. Depending on the microarchitecture of the CPU, overclocking (assuming it is possible) might even be detrimental to performance in certain cases; for example, maybe increasing the CPU speed beyond a certain point will result in one or more of the cores shutting off (which would be bad if the software you are running is optimized for multicore scenarios, but might be good if your software only uses one core).
Lastly, when it comes to emulators or apps/games in general, obviously beefier hardware helps and is even sometimes a necessity depending on what it being emulated or rum. However, with that being said, optimizations and programming skill can go a long way towards creating applications that perform well. In addition, it must be said that the Vita, like other dedicated consoles, offers developers a non-fractured base (everyone with a Vita gets a system with the same microarchitecture) and deeper hardware access, which leaves plenty of room for optimization. Furthermore, many of the emulators that are running natively on the Vita are poorly optimized and so there is a lot of room for overall improvement. As the OP said maybe a slight clock speed bump will make a small difference on some games and emulators (I very much doubt a huge overclock would be achievable or practically desirable even if it was possible), but I feel that at this point the focus should be on optimization. After all look at what was made possible on the much less capable PSP through optimizations.
D
Deleted User
Guest
you are forgetting that henkaku in fact is a kernel exploit... so it is possible.
It is but it is not currently released in open form for anyone to tinker with it (aka make other CFW or expand Henkaku features like version spoofing and so on).
Once race will be finished i expect uptake on CFW featureset as people will be able to tinker with it properly.
again, please read my previous comments. I never once commented about even overclocking it beyond 1Ghz, I mentioned speed like 500-700mhz, which based on the heat generated from A15, they should be able to sustain a long time before heat throttle comes in. A15 is a very powerful chip back then and is used on flagship chips like Exynos 5, Tegra 4, so I have no doubt on their performance. I recall those phones are able to hold some rather good clockspeed ( 1ghz or above ) over long duration during gaming while getting the heat rather well maintained in mobile phone. With that in mind, I come out with this thought that running on 700mhz would be theoretically okay for Vita. Snapdragon 810 is a bad example, given that it is known to overheat easily with high power draw compared to earlier architectures and manufacturers have to resort to aggressively throttle it to maintain an acceptable temperature.
Yeah, those comment gets to me too at times, especially on mobile phone in the last couple of years where end users just blindly chasing the number of cores instead of looking at how the whole architecture is designed to give higher performance. On the new 3DS, I recall nintendo did say that they increased the CPU speed by a rather huge margin, i recall it was around 30%, but might be wrong , and the performance boost is mainly gained from there,I do not follow the chip architecture on both console, but I would assume that they are using different chips, 3DS CPU being a newer architecture compared to PSP. Different processor have difference performance per clock, and naturally, 3DS being a later product would have a much greater performance per clock. It's like comparing a recent i5 with 4 cores with a first gen Quad Core processor, given the same clock speed and turbo boost disable, it would still be very obvious that i5 would still win hands down.
that being said, i agree to you that there are more to just clockspeed boosting, software optimization is equally important. I am a software programmer myself, so I know what kind of boost software optimization could bring. Perhaps in the future there will be good modders that can cook up CFW that gives performance boost purely based on software optimization, but when it comes to performance, I believe the more the merrier, if software optimization can already give us some boost, imagine what it would do with an increase of maybe 20-30% clock speed?
Well if it will be possible I am curious to see whether some games might improve their performance, namely Jak and Daxter trilogy which works at like 15-20 fps
What a disgrace it was to release the game in such a form. Just goes to show you how important optimisation is.
I think it's the only time that a game is not even worth it game stop trade in value.
Whoa wait what?? Does it do this automatically if you turn off wifi or do you need to enable it, or only with certain games?
Developers can turn on 444 mhz, but they also need to make the game turn off the Wifi themselves. Some games run at 444 mhz, but don't turn it off and you can actually see the performance gap between them. The game Legend of Dark Witch is a notable one where downloading something makes the game crawl.
Last edited by Cinnamon,
Similar threads
- Question
