Related
Recently, I finished finally modifying the xperia x10 that I have and one of the greatest improvements that those guys achieved was getting a lot more processing power by enabling JIT.
So naturally after seeing it work miracles on the x10, I went to look for it on my captivate and so far have come up with nothing anywhere. I saw some discussion in the past about it but nothing beyond.
Hopefully somebody can enlighten and if such a thing does seriously want to be worked on.... Here's a thread!
Sent from my GT-I9000 using Tapatalk
FroYo (2.2) and later have Just-In-Time compilation (JIT) out of the box. There are probably a dozen FroYo ROMs on the front page alone, so pick one have fun!
From what I've read, it's been a part of the ROMs since 2.1 but it needs to be enabled. Is this the case here or am I missing something?
Sent from my GT-I9000 using Tapatalk
kr3w1337 said:
From what I've read, it's been a part of the ROMs since 2.1 but it needs to be enabled. Is this the case here or am I missing something?
Sent from my GT-I9000 using Tapatalk
Click to expand...
Click to collapse
That is one of the biggest parts of froyo is that it enables JIT...so if you run a rom that has froyo its enabled unless I'm mistaken...
Sent from my GT-I9000 using XDA App
JIT is enabled by default in all Froyo ROM's. You can check the build.prop for ****s and giggles though.
The Galaxy S won't show the kinds of scores that Qualcomm based devices will, even with JIT enabled. Qualcomm included 128-bit SIMD Floating Point extensions with Snapdragon, while the Hummingbird only has 64-bit extensions.
yes with jit we get a 60-70% improvement but a qualcom gets 300%+ in floating point operation. in modern computers there are many other factors though and quardrant cpu scores are still very high for our chip (if you use the pay version you can see the score breakdown). so dont let the linpack scores discorage you. ive gotten as high as 18.2 in linpack with some overclocking though, which isn't bad. it's not the 50+ in the qualcom phones with some mods but not bad.
So that's what the story is, thanks guys! Wondering if things could be improved beyond overclocking...
I know it can be but don't know how. Some guys are getting 25+ in linpack on there website. So there is something else holding us back a bit.
Sent from my SAMSUNG-SGH-I897 using Tapatalk
Dani897 said:
I know it can be but don't know how. Some guys are getting 25+ in linpack on there website. So there is something else holding us back a bit.
Sent from my SAMSUNG-SGH-I897 using Tapatalk
Click to expand...
Click to collapse
You talking about xperia x10? really 25 on linpack? I dunno if that is even possible to OC to astronomical speeds. They run on older hardware and I think most are still on 2.1.....
All Android ROMs have a JIT compiler, it literally is what compiles all the Java on the fly. Newer versions are optimized for better performance.
So the better question is do the 2.2 Froyo ROMs have the latest JIT compiler version available or does the Nexus S have a newer more performant version we can steal. More than likely it will not work with 2.2 since the Nexus S is using 2.3.
From everything I have heard the newer JIT comp versions are optimized for the Snapdragon chipset more than anything. Which doesn't do us much good.
2.2+ have the JIT. Prior to 2.2, all programs ran entirely as interpreted bytecode on an isolated virtual machine. In 2.2+, the JIT translates the most cpu "heavy" bytecode down to native instructions during execution, stores it in cache, then runs it natively on the processor in a protected mode. Dig the video below, it's an hour long but the functionality of the JIT is explained in the first 15 minutes.
^Edited the above to accurately describe the function of the JIT
http://www.youtube.com/watch?v=Ls0tM-c4Vfo There you go, dudes. JIT Demystified.
modest_mandroid said:
2.2+ have the JIT. Prior to 2.2, all programs ran as bytecode on an isolated virtual machine. In 2.2+, the JIT translates the bytecode down to native instructions just before execution, then runs it natively on the processor in a protected mode. Could be wrong, but to my understanding, that is the major difference between 2.1- and 2.2+.
http://www.youtube.com/watch?v=Ls0tM-c4Vfo There you go, dudes. JIT Demystified.
Click to expand...
Click to collapse
This is what I saw and know too. However, I believe that the compiler that we are using is rather old and has some room for improvement. The nexus S compiler working on the captivate is a possibility that could become true after the 2.3 port is finished, provided that the developers look into it.
Sent from my GT-I9000 using Tapatalk
Linpack v1.2.7 was released and I found a 4MFLOP increase without changing anything on my phone. Anyone else see a difference in their score? The update stats it only added multi-threading support but I believe something else changed.
Same for me as well. Must have changed the testing method. Went from 14ish with AAi3 to 17.5ish
abhaxus said:
Same for me as well. Must have changed the testing method. Went from 14ish with AAi3 to 17.5ish
Click to expand...
Click to collapse
I'm getting 17.5 on Bonsai 5 also now. Here is a note from the Market description of the update-
"Note: Your MFLOPS rating will increase in this version due to updated libraries and methods. Shows Android is improving! Newly added the ability to fully test multi-core processors with the use of multi-threading. Compare speeds from a single and multi-thread runs. See how well your multi-core device works under android."
So not only does it state multi-core support, but also new libraries that increase the scores on all devices. Kind of confusing for a benchmark app to do this as it makes all the old scores obsolete.
Thank you very much for explaining why the scores increased. I should of read it all before upgrading.
Sent from my SPH-D700 using Tapatalk
I got a 23 that's the highest I've gotten
Sent from my SPH-D700 using XDA App
my Atrix was flying before i had to send it in for warranty work a few weeks back. (digitizer) Antutu scores around 7000. the rest of the benchmark tests where pretty high also. When it was returned i immediately unlocked the bootloader and rooted it. Now my scores from Antutu are 4500 at best. I am using the exact same rom (Neutrino), and kernel (Faux123) as before. I have tried other roms (cm7 nightlies) reinstalled and full fastboot wipe many times but still the same outcome. The only difference i can see is that i used an different method to unlock the bootloader and root the second time around. Do you think the some of the permission where affected differently the second time around. Am i just overlooking something?
HELP!!!!!!
thks
Is the phone actually slower? Or just scoring lower scores?
Sent from my HTC Inspire 4G using XDA
Benchmarks mean nothing
Sent from my MB860 using xda premium
Alcapone263 said:
Benchmarks mean nothing
Sent from my MB860 using xda premium
Click to expand...
Click to collapse
That was going to be my next statement.
Sent from my HTC Inspire 4G using XDA
the phone actually feels slower. more laggy when it comes to videos, games, scrolling through menus. my fps scores are all lower across the board in any benckmark. yes i know they mean nothing but when the scores are that drastically effected there is an issue i believe.
gentledroid17 said:
the phone actually feels slower. more laggy when it comes to videos, games, scrolling through menus. my fps scores are all lower across the board in any benckmark. yes i know they mean nothing but when the scores are that drastically effected there is an issue i believe.
Click to expand...
Click to collapse
I have heard that refurbished phones USUALLY don't perform as well as the brand new ones do. Here is my example, a refurbished Atrix I read from some users can't even sometimes boot a custom kernel when the custom kernel is around 1.45 to 1.6. For example the one kernel that the dev clemsyn. I saw a FEW mention that when they said they had a refurbished Atrix.
Sent from my Atrix with XDA Premium
gentledroid17 said:
my Atrix was flying before i had to send it in for warranty work a few weeks back. (digitizer) Antutu scores around 7000. the rest of the benchmark tests where pretty high also. When it was returned i immediately unlocked the bootloader and rooted it. Now my scores from Antutu are 4500 at best. I am using the exact same rom (Neutrino), and kernel (Faux123) as before. I have tried other roms (cm7 nightlies) reinstalled and full fastboot wipe many times but still the same outcome. The only difference i can see is that i used an different method to unlock the bootloader and root the second time around. Do you think the some of the permission where affected differently the second time around. Am i just overlooking something?
HELP!!!!!!
thks
Click to expand...
Click to collapse
There is a "Boost.sh" script included with Neutrino (I believe in /system/etc) that Notorious recommends be run when you get that "not so fresh" feeling. It is a defrag tool that can be assigned to a widget and should be run every so often.
Also, you will want to see if "Nitrocharger.sh" in the same directory is running as Root at Boot. I have seen big improvements by implementing this.
Both of these scripts came with the ROM, so hopefully you have fast access to them. Good luck to you!
thanks for the info fellow xdaers,,,,,
Voelker45 said:
I have heard that refurbished phones USUALLY don't perform as well as the brand new ones do. Here is my example, a refurbished Atrix I read from some users can't even sometimes boot a custom kernel when the custom kernel is around 1.45 to 1.6. For example the one kernel that the dev clemsyn. I saw a FEW mention that when they said they had a refurbished Atrix.
Sent from my Atrix with XDA Premium
Click to expand...
Click to collapse
Actually I had the opposite experience. My original phone wouldn't boot with the clemsyn kernel. My digitizer bombed and I received a refurbished unit from repairs. The new unit runs the clemsyn kernel just fine. So I guess it's just the luck of the draw...
Sent from my MB860 using xda premium
tobnddl said:
There is a "Boost.sh" script included with Neutrino (I believe in /system/etc) that Notorious recommends be run when you get that "not so fresh" feeling. It is a defrag tool that can be assigned to a widget and should be run every so often.
Also, you will want to see if "Nitrocharger.sh" in the same directory is running as Root at Boot. I have seen big improvements by implementing this.
Both of these scripts came with the ROM, so hopefully you have fast access to them. Good luck to you!
Click to expand...
Click to collapse
gentledroid17 said:
the phone actually feels slower. more laggy when it comes to videos, games, scrolling through menus. my fps scores are all lower across the board in any benckmark. yes i know they mean nothing but when the scores are that drastically effected there is an issue i believe.
Click to expand...
Click to collapse
As tob said, those "so verry little script" can makes big differences sometimes
I would like to add my personal toughts, for me, the ROM actually scoring higher and higher and feel a litle bit more freshier each week past within the ROM... so my thoughts would be, the cache partition, the more longer the rom is implemented, the more score you'll get cause the cache slowly become bigger and bigger giving you full power to open something tough!
As for the battery, from my experiences, you can't really judge a battery life while you just flashed a rom, you got to wait until the first week to see if there a big changes between the other ROM you came from ... so my guess would be, those scripts and the longer the ROM is implemented on your phone, the more fast and stable it will get, and you will feel the power slowly coming out again!
My bad if I'm completely wrong, but it is my guess!
Hood luck with your "new" phone
i brought up some old Antutu scores and notice that my ram. cpu integar and floating point scores are extremely low. i have tried multiple kernels but still no change. now i am even more confused......doesn't take much. could it be that my cpu is locked so i can not overclock. it kinda feels like its only utilizing one core.
after weeks of exhaustive searching i finally found a fix that worked.
http://forum.xda-developers.com/showthread.php?t=1131649
pdsfix.zip
now my antutu scores are back up to around 7000. less lag in app drawer and opening /closing certain app. its back to silky smooth yeehaw.
Hey XDA friends,
I was testing my One today with some intense gaming and benchmarking. I noticed that Real Racing 3 was slowing down after some time of gaming. So I ran Antutu before and after my gaming session. My score before: 24567, my score after 20 minutes of Real Racing 3: 19945 !
Especially the GPU score was significantly lower in the second test. It seems that the One is throttling speed when reaching defined temperatures. There is a thermald.conf in /system/etc , but impossible for me to read out. A modified thermald.conf or a custom kernel should sort that out!
Sent from my HTC One using xda app-developers app
more2come said:
Hey XDA friends,
I was testing my One today with some intense gaming and benchmarking. I noticed that Real Racing 3 was slowing down after some time of gaming. So I ran Antutu before and after my gaming session. My score before: 24567, my score after 20 minutes of Real Racing 3: 19945 !
Especially the GPU score was significantly lower in the second test. It seems that the One is throttling speed when reaching defined temperatures. There is a thermald.conf in /system/etc , but impossible for me to read out. A modified thermald.conf or a custom kernel should sort that out!
Sent from my HTC One using xda app-developers app
Click to expand...
Click to collapse
Mh.. we had a guys that ran a lot of taiji basemark runs (loop) and had no problem with throttling. Maybe Taiji is gpu only and real racing 2 needs gpu and cpu a lot. A bit worrying though.
Thermal throttling really is a big problem for S4 Pro. So S600 has to suffer it also, although HTC made a very thin copper cover inside for reducing the heat.
shox22 said:
Mh.. we had a guys that ran a lot of taiji basemark runs (loop) and had no problem with throttling. Maybe Taiji is gpu only and real racing 2 needs gpu and cpu a lot. A bit worrying though.
Click to expand...
Click to collapse
GPU tests don't make a significant stress on the system. The thermal throttling happens mostly when both CPU and GPU reaching the limit.
That's too bad... :-/
My Nexus 4 really suffered from thermal throttling, too. But a custom kernel by Franco with lowered voltages and a newly written thermal handling routine sorted that out, the SoC didn't become too hot and didn't throttle at all!
Can someone try running Stability Test (cpu+gpu test) to see at what temperature is throtled?
Re: HTC Flagship 2013 - HTC One - Qualcomm® Snapdragon™ 600 - Mega Information Thread
May be an isolated issue and not with all devices?
more2come said:
Hey XDA friends,
I was testing my One today with some intense gaming and benchmarking. I noticed that Real Racing 3 was slowing down after some time of gaming. So I ran Antutu before and after my gaming session. My score before: 24567, my score after 20 minutes of Real Racing 3: 19945 !
Especially the GPU score was significantly lower in the second test. It seems that the One is throttling speed when reaching defined temperatures. There is a thermald.conf in /system/etc , but impossible for me to read out. A modified thermald.conf or a custom kernel should sort that out!
Sent from my HTC One using xda app-developers app
Click to expand...
Click to collapse
Sent from my Windows 8 device using Board Express
I can hardly believe that this is an isolated issue! I'm pretty sure that the S600 has the same temperature handling routines as the S4 Pro. The presence of a thermald.conf in /system/etc/ shows that there are defined parameters for SoC throttling to protect the device.
If you're willing to test that, just give Real Racing 3 a go for about 30 minutes and run Antutu directly afterwards. Or run stability test for longer periods of time.
If this is the case, you would have thought Qualcomm would have at least attempted to sort out the thermal issue as seen in some S4 Pro handsets!.
more2come said:
I can hardly believe that this is an isolated issue! I'm pretty sure that the S600 has the same temperature handling routines as the S4 Pro. The presence of a thermald.conf in /system/etc/ shows that there are defined parameters for SoC throttling to protect the device.
If you're willing to test that, just give Real Racing 3 a go for about 30 minutes and run Antutu directly afterwards. Or run stability test for longer periods of time.
Click to expand...
Click to collapse
I did what did you ask there is not really difference in Antutu result, before and after, both above 22K.
I played around 25 minutes of RR3
Sent from my HTC One using Tapatalk 2
@MrKaon
Are you on some custom ROM?
No m8 Stock
Sent from my HTC One using Tapatalk 2
MrKaon said:
No m8 Stock
Sent from my HTC One using Tapatalk 2
Click to expand...
Click to collapse
Thanks, that and your signature were very helpful! My stock firmware was from O2 Germany which is reportedly older than the most recent stock EU firmware. I flashed mike's Revolution ROM right now which has a newer firmware base and possibly a newer kernel. Thermal throttling doesn't happen on Revolution HD!
more2come said:
Thanks, that and your signature were very helpful! My stock firmware was from O2 Germany which is reportedly older than the most recent stock EU firmware. I flashed mike's Revolution ROM right now which has a newer firmware base and possibly a newer kernel. Thermal throttling doesn't happen on Revolution HD!
Click to expand...
Click to collapse
Well that's reckless and dangerous.
There is thermal throttling for a reason, and it's set very aggressively for a reason.
Hunt3r.j2 said:
Well that's reckless and dangerous.
There is thermal throttling for a reason, and it's set very aggressively for a reason.
Click to expand...
Click to collapse
Thermal throttling didnt steped in - but that doesnt mean that thermal throlltling isnt there. The change could come from the update to 1.28 (new driver for example). The same happened on the one x last year. There was an update that significantly reduced heat.
shox22 said:
Thermal throttling didnt steped in - but that doesnt mean that thermal throlltling isnt there. The change could come from the update to 1.28 (new driver for example). The same happened on the one x last year. There was an update that significantly reduced heat.
Click to expand...
Click to collapse
It gets a little less hot on the backside with the newer firmware base, I can confirm that.
it gets hot, it throttles. it's a safety measure. why do u complain?
so u want your phone to breach the heat threshold then the chip dies and then u're gonna say htc one sucks becoz components die?
shiningarmor said:
it gets hot, it throttles. it's a safety measure. why do u complain?
so u want your phone to breach the heat threshold then the chip dies and then u're gonna say htc one sucks becoz components die?
Click to expand...
Click to collapse
Of course it's a safety measure, but in my opinion a smartphone shouldn't throttle due to normal usage. If the circumstances are abnormal (very hot summer temperatures, for example), throttling is okay and I'd welcome that to protect my valuable HTC. But while playing a game?
But there's no further reason for me to complain, because the newest firmware runs cool enough without throttling.
so in the latest firmware its not throttling? can you add this info to your OP?
What is the latest firmware? Is there a published changelog and do I need to go back to stock to get at it?
Sent from my Tricked out HTC One via xda-developers application
Tetsumi06 said:
What is the latest firmware? Is there a published changelog and do I need to go back to stock to get at it?
Sent from my Tricked out HTC One via xda-developers application
Click to expand...
Click to collapse
latest should be 1.29
Is there a way to have individual core Control clock speeds and governor if possible I'd prefer an app for obvious reasons (easier) I'm on cm10.1 5/9 nightlies
Sent from my Nexus 7 using Tapatalk 2
Some of the rate governors (not all of them) let you select the maximum number of cores allowed to be online. Depends on the kernel, but in principle you can use Trickster Mod. While clocking on the Tegra 3 is quite flexible, I believe it is not possible to have separate G cores operating simultaneously with different clock rates.
That's lame the subject came up because I have it working on my Atrix HD AT&T but I think I'll try another kernel
Sent from my Nexus 7 using Tapatalk HD
Franzferdinan51 said:
That's lame the subject came up because I have it working on my Atrix HD AT&T but I think I'll try another kernel
Sent from my Nexus 7 using Tapatalk HD
Click to expand...
Click to collapse
What's the point of it?
bftb0 said:
What's the point of it?
Click to expand...
Click to collapse
Extra battery but more speed with a kind of stepping stone per say look at it like this using my dual core 1.5 ghz atrix hd as an example
Sleep one core and you run single core with lots of lag
But with this method you can under clock core 0 1ghz to and change the government to interactive use the second core as something to the n7 companion core take it way down farther let's say 600mhz with on demand or possibly conservative governor. That way the second core would come on in times of lag for a small push to end lag spikes and like I said works on my atrix quite well
Sent from my Nexus 7 using Tapatalk HD
I think I would expect what you are describing to exhibit strange (pathological) behavior unless all of the rate governors are re-coded to collect their heuristics partitioned by processor thread affinity.
Does this also mean the 2nd processor is never off-lined? (They can drain a lot of juice even when underclocked due to static power dissipation issues, so it makes me wonder if the power savings is real)
Is the kernel development work for that device (Atrix HD) described anywhere by the implementer(s)?
A couple I couldn't point you to a definitive answer as I'm not a dev though it's defiantly someplace here on the forums even a kernel to look at
Sent from my Nexus 7 using Tapatalk HD
Franzferdinan51 said:
A couple I couldn't point you to a definitive answer as I'm not a dev though it's defiantly someplace here on the forums even a kernel to look at
Click to expand...
Click to collapse
Are you talking about that script-ware by smokin1337? If so, it seems to (try to) work by forcing the second core to be on-line at all times, and then changing rate governors on a per cpu basis, not in the kernel but by continuously writing to each cpu entry in sysfs.
I peeked over in the Atrix HD forum and it seems it doesn't even have any working custom kernels yet...
http://forum.xda-developers.com/showthread.php?p=40253686
That's the only kernel to my knowledge
Sent from my Nexus 7 using Tapatalk HD
@Franzferdinan51
Maybe you could throw me a bone - what exactly is it that you are using on your Atrix HD that does this? (Is it baked in to somebody's ROM, or a separate flashable patch)?
I *did* go searching over in the Atrix HD forum rather extensively.
Downloaded Codex01's "CM10.1PreformanceEnhancements-3.0.1" and looked in there - this doesn't do what you say.
Downloaded tcf38012's popcorn kernel and unpacked it and poked around - it also doesn't do what you say (lots of other tweaks tho).
Found a mention of something similar in posts by skeevydude. Downloaded smokin1337's "CPU Editor" for snapdragon - it was mentioned in passing in the Atrix CM 10.1 thread.
Am I just looking at the wrong things?
Anyway, everything that I've found so far that looks close to what you are describing writes control information to stuff in /sys/devices/system/cpu{/cpu0|/cpu1}/*
For what it is worth, that same (sysfs) stuff does exist in various N7 kernels - for instance, per-cpu entries for min/max frequencies and rate governors in /sys/devices/system/cpu/cpu{0|1|2|3}/cpufreq/.
So, maybe what I said first was wrong. Maybe the right answer should have been "kinda - maybe - sorta". I would have to understand the PLL schemes that different kernels use a lot better than I do to be definitive.
But I am still a bit skeptical that it actually produces the result that it claims - saving battery life by forcing two cores to be online at all times... without also affecting performance. And the part about two independent control loops affecting each other in pathological ways remains open as well (threads running on the other core with a different rate governor affect the measurement of the recent system load averages used by the second rate governor - and vice versa).
It would be useful to have a decent and repeatable way to benchmark interactivity - the first-person reports of "this is really smooth" or "lags badly" are always completely subjective and non-repeatable, so it is hard to know who to really believe when it comes to reports about this stuff.
cheers