Hermes WM Roms Benchmarks
My Hermes is dead.. this isnt going to get updated again, im so sorry
One of the MOST common questions we get, is, what's the lites rom? And what's the fastest rom?
How to make your rom faster
A new program called WMRegOptimizer, which you can get here, has had noticed improvements on the speed of my phone and others.
For more detailed info and a few helpfull recommendations.
My Methods:
I take the rom, fresh from the chef, and flash it, that's it, no UC, no changes, nothing!
I run the tests unplugged, and before connecting to activesync.
All i install is my battery drain tool, and my benchmarking tool.
I dont change the Pagepool, or the cache settings, for the main benchmarks.
Below there will also be benchmarks of the rom i use, and other members of XDA, post yours up as well!
Also, attached you will find a XML you can import with Advanced config, or use in your UC, this is cache tweaks, and i will post benchmarks to show how they help.
I know that My benchmark suit isnt the best around, i chucked it together in 5min, but, as im using the same on all roms you can compare without worrying to much!
If there was any "odd" readings, aka something much higher/lower, I re benchmarked, re flashed, until I was sure its not an error I can do anything about.
About each section :
Rom name: name of the rom
Rom build: The build of the rom
idle power consumption (mA): This is the power it was draining in mA (milliamp) sitting in the power consumption tool with the backlight (at rom default), the device was not allowed to turn off ect. And i left it for 2-5min to average.
Integer: How quickly the phone performed Integer calucations, see here for more info
floating point: How quickly the phone performed floating point calucations, see here for more info
RAM Access: How quickly the phone could access the RAM
Draw Bitmaps: How quickly the phone could redraw bitmaps on the screen (graphical performance)
Main Storage read/write: How quickly the phone could read/write to the main storage.
Higher is better apart from for idle power consumption, where lower is better.
Power consumption is VERY MUCH affected by screen brightness, I should have standardised this.. but it still give an idea..
My conclusions
RBS roms are faster than Non BS
Compressed roms are slower and use more power in idle for the most part
Some roms are faster but use more idle power, and different roms have there strengths and weaknesses!
.zip has all the benchmarks so far, along with graphs.
Pagepoll tests
I decided to take 3 roms and test 6v8meg PP, the results are in the .zip
For the most part, the 8meg PP roms where faster, but has less ram (duh). I would do more tests, but it would take far to much time..
Graphs
These are out of date compared to the .zip, I will update them asap.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
There ALL in the .zip, they would take up to much room to post here.
Performance XML
This file can be run with UC, or imported with advanced config.
What this does is fine tune the cache sizes to what I think gives the best performance.
Updated on 25th August
Added Power management tweaks to the XML
Updated the Cache settings.
Added tweak to remove the delay when taking pictures.
Examples:
CRCv12.9 BS naked 8meg PP pre XML benchmarks. (See that the higher pp is faster than the 6meg one in the main benchmark tests)
Integer: 181.9167
Floating Point: 5.779
Draw Bitmaps: 710
RAM Access: 1620
Main storage (write): 710.62
Main Storage (read): 3690.09
CRCv12.9 BS naked 8meg PP post XML benchmarks.
Integer:184.3587
Floating Point: 5.783
RAM Access1632
Draw Bitmaps: 726
Main storage (write)710.37
Main Storage (read)3685.66
Performance settings if you dont want to use the XML:
File system cache: Enabled
File system cache size: 2megs
File system filter cache: 8192 sectors
Glyph Cache: 128KB
What to benchmark your own rom? See this post for the program I use!
Notes: None of the tweaks in the XML is my work, apart from making it and posting it here, im just trying to make it easier for the newbies to get the most from there device!
WM 6 roms
This thread here by KarhU has benchmarks for most WM6 roms
Todo:
*Add links to rom thread in the spreadsheet.
*Work out why there is the MASSIVE increase in main memory speed on CRCv12.7 and roms based on it. (seems to be some due to some tweaks in the rom!)
If this has helped you, or you like what I do, maybe consider a small donation, anything is very gratefully received!
I'm shaking
In my chair, waiting for the results, ..... half an hour too long, get your a into g
zocster said:
In my chair, waiting for the results, ..... half an hour too long, get your a into g
Click to expand...
Click to collapse
It takes 10min per rom And i keep having to text/call mid power usage test ;/
Bravo
Awesome work, ...
Excellent job, thanks a lot !!
That all the current roms done, Ill try to make it easer for you to see what everything means tomorrow, right now I need to get working, get my Hermes back to how I like it, and do some work
Any feedback? Any roms you want me to do? any graphs? Just say!
Impressive..simply impressive..
blackout128 said:
Impressive..simply impressive..
Click to expand...
Click to collapse
Thanks you!
Now, to get the ball rolling, lets see your benchmarks!
Heres mine: CRCV12.9BC-naked 8meg PP
Integer: 181.1740
Floating point: 5.773
Ram Access: 1637
Draw bitmap: 721
Main Storage (write): 708.24
Main Storage (read): 3683.45
veyka said:
Thanks you!
Now, to get the ball rolling, lets see your benchmarks!
Heres mine: CRCV12.9BC-naked 8meg PP
Integer: 181.1740
Floating point: 5.773
Ram Access: 1637
Draw bitmap: 721
Main Storage (write): 708.24
Main Storage (read): 3683.45
Click to expand...
Click to collapse
Mine fully loaded, with MSN pushing mail, and activesync on 2003 OWA sync.
integer: 178.6745
floating pt: 5.751
RAM Access: 1600
Draw bitmaps: 565
Main Storage Write: 614.46
Main Storage Read 3723.64
Heres mine: PDACorner NRBSN V20 8 Meg PP
Integer: 176.8177
Floating point: 5.746
Ram Access:1626
Draw bitmap: 716
Main Storage (write):631.45
Main Storage (read): 3453.63
@ veyka!
idle power consumption higher is better or lower is better ?
dvha79_ddt said:
Dear peyka!
idle power consumption higher is better or lower is better ?
Click to expand...
Click to collapse
Lower is better! Wooops! Ill correct that in my first post. Thanks
veyka said:
Lower is better! Wooops! Ill correct that in my first post. Thanks
Click to expand...
Click to collapse
Thank you for your fast reply, This topic is real mean.
Good luck!
dvha79_ddt said:
Thank you for your fast reply, This topic is real mean.
Good luck!
Click to expand...
Click to collapse
Thanks again! Im just benchmarking another 4roms! update asap.
5 roms added! New better graphs planned for later today!
Awesome work!
Keep it up!
Any Schap's 4.40 Beta benchmark??
Nico
thank, man - you do a nice work
Thanks man very useful, I'm still on WinMo6.0 deciding where to go.
colombia7 said:
Any Schap's 4.40 Beta benchmark??
Nico
Click to expand...
Click to collapse
I can do! Ill look into getting a non locked copy!
And thanks everyone
I ran out of steam last night But ill finish off the updates for today.
Related
Here are some tests I have done on BLACK 2.5
File System Memory Allocation: the more memory WM6 allocates, the faster it will run.
Font Cache: the more memory allocation, a faster UI is possible.
Black 2.5 CLEAN Install - FREE RAM on Soft Reset: ~32MB
File System Memory Allocation: Auto
Font Cache: 8192 bytes
Customizations for you to try:
Buzzing - On "SPEED" WM6 (FREE RAM: ~24MB)
File System Memory Allocation: 8mb
Font Cache: 32768 bytes
Performance - Faster WM6 (FREE RAM: ~28MB)
File System Memory Allocation: 4mb
Font Cache: 32768 bytes
Memory - More Free RAM trading off Speed (FREE RAM: ~ 32MB)
File System Memory Allocation: close to zero
Font Cache: 2000 bytes
Going back to the Defaul BLACK 2.5 setting:
Default BLACK - Let WM6 decide - (FREE RAM: mixed results depending on usage)
File System Memory Allocation: Auto (determined by WM6)
Font Cache: 8192 bytes
Ensure you SOFT RESET after installing one of the cabs.
Also, if moving from one CAB setting to another e.g. Memory --> Buzzing, then uninstall Memory, then install Buzzing and soft reset.
Post your results here in this thread.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Only for Black 2.5 or can we all use it?
PAPPL said:
Only for Black 2.5 or can we all use it?
Click to expand...
Click to collapse
no probs, all can use it, but im not sure what the default settings are on the other ROMS for you to revert back to.
hello !
what does it really do ? can this harm the machine ? what about battery (faster = drain of batt) ?
Thx in advance
DR400 said:
hello !
what does it really do ? can this harm the machine ? what about battery (faster = drain of batt) ?
Thx in advance
Click to expand...
Click to collapse
It does not harm the hermes, it just allocates more or less memory (leaving less or more usable) to the file system.
Not sure about battery implications but I beleive it will be the same.
ok I trust you on this one ... i'm gonna try ... let you know about drain of battery (if any) in line with my personnal usage.
just installed ... it looks to me like as fast as black 2.0 ... by the way I was a bit disappointed by 2.5 in comparison with 2.0. Execution of softwares seemed to be faster in 2.0. Using the "buzzing" effect seems to equal in a way the reactiveness of 2.0
Just first thoughts ...
thanks jasjamming!
i've tried all, and i prefer 'performance'. also, i've change the close application configuration. a tap exits the application, a longer tap closes it. this gives me more control.
Cool, i'll have to play with these! When you allocate more ram to windows, does it come off the Program or Storage space?
As a side note, i havn't noticed Black 2.5 to be any slower than 2.0...
Nice job JJ - could this be made into a utility?
Hey JJ, nice work dude - will give those a try over the weekend (on call until Friday, so phone has to stay in one piece until then!!)
As a matter of interest, if those CAB files are "just registry settings" (haven't looked, so this is just speculation!), could a utility be written to allow the same values to be adjusted by the user? The system memory and font cache could be controlled by a slider, and the options in the CAB files could be auto-selected from a drop-down list; once any change is made, the user gets the chance to Apply or Cancel and if they tap Apply, the phone is soft-reset....
****, I've just written a program spec in three minutes
Just a thought - obviously, if the CAB files do some things that can't easily be handled at the UI level, this is a non-starter
......reckon Walshie could knock this together between tinnies
What say ye?
Mark.
Oh yes, I like buzzing......
JamesAub said:
As a side note, i havn't noticed Black 2.5 to be any slower than 2.0...
Click to expand...
Click to collapse
well just a personal opinion ... maybe switching from WM5 to black 2.0 was such that the gap between black 2.0 and 2.5 was not what expected ...
black 2.0 was damn fast for me ...
black 2.5 is still fast but ... well ... could be better for some applications I use.
but thx JJ for the work done ... those remarks are just for adding my humble stone to the mountain of work you have done so far
installed buzzing, hoping for some improvements! thanks for this
On older WM6 versions changing cache size made the device unstable
Did you notice any stability changes? does your device freeze?
by default all cache is set to auto and WM6 allocates more and more to the cache(something many users thought was memory leakedge).
I think 8MB is more then enough, the FONT cache should be on 32K(from my tests on the older WM5).
JASJAMMING, your the BEST, thank you for all the efforts!
so guys, Did you notice any stability changes?
i think it's too early to judge on stability. let us run for days and comment later. so far so good for me.
I'm on the Performance setting - can notice a slight difference....
Shot JJ - Pro 2.5 is going AWESOME so far...
Just installed buzzing and my 8525 is definitely more snappy.
Did a quick test with multiple apps running (Messaging, WMP -playing music, IE, and Live Search open) and everything is running smoothly. Although the 8 mb takes a big chunk out of your program memory.
Looks like more great work JJ.
yums
jasjamming said:
no probs, all can use it, but im not sure what the default settings are on the other ROMS for you to revert back to.
Click to expand...
Click to collapse
JJ,
This is excellent work Bro.
Thanks,
DR400 said:
just installed ... it looks to me like as fast as black 2.0 ... by the way I was a bit disappointed by 2.5 in comparison with 2.0. Execution of softwares seemed to be faster in 2.0. Using the "buzzing" effect seems to equal in a way the reactiveness of 2.0
Just first thoughts ...
Click to expand...
Click to collapse
Have to admit I was thinking the same, 2.0 flew and 2.5 was seemingly a little slower. I didn't want to post my thoughts without giving it a chance but have been using since release day.
I have just installed buzzing and early indicators are good, I will report back on stability/speed in a few days.
Thanks JJ.
Almost a month and no ROMs for 2008 yet? I've waited enough and this made me cook my own. Found that there are some OS 5.2.2000 builds for Artemis in different languages but as far as I and other senior members know, MS hasn't released such kind of a build and its just a fake. Even if it is true, there are no die-hard improvements in it. So I started off with a version that is stable and fast for Professional use. Version 5.2.1933
Removed every possible file that most wont require or the ones which can be installed anytime via cabs. No HTC Home plug-in, Touch Bio Scroll, Cube and all those crap which were bought out to market HTC' newer devices and their fantasizing period is over. Even Transcriber, MMS and other stuffs are removed. Maybe if you guys need them, I shall build another version.
Added:
-1 Calc lite -replacing default calculator
-Net CF 3.5 -replacing Net CF 2.0
-Battery Status plug-in
-Advanced Configuration Tool by Schap
-Chess -removed other games
-CommManager -Touch version
-H Button
-Oxios Memory
-PS ShutXP
-HTC Album
-Resco File Explorer(trial)
-PIM Backup
-Pocket RAR
-Pocket xPDF
-WM5New Menu
-Task Manager
-Screen Capture
-Key Board Skin
-iBliss Phone Skin
Registry Hacks
-ClearType Enabled
-File System and other cache settings tweaked for better performance
-Email, attachments, Temp IE files all moved to Storage Card
-Security Warning Disabled
Final Result:
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
The perfect fast and stable ROM I waited for! The device works much faster now and its completely clean without any crap. It took me off around 80 hours to build this ROM. Because I was not a cook before and so learnt everything from scratch. Sat 10hours for 8 days continuously, made ROMs that never booted, almost bricked my device and finally made it. Big thanks to the kitchen makers and other ROM masters here in xda-developers.
Download
Reserving this thread..
Regards,
Carty..
Wow Carty,
You got my Vote!
One of the best roms for my artemis!
Wondering what Tom's have to say about this
how much program and Storage memory has your ROM please?
Any screenshots of this Rom will be appreciated.
Yeah... We want screenshots! And info about memory. What about Office..?
Hello
Congratulations for your job . I have Artemis Touch 3.01 . Can somebody to make an comparison between Artemis Touch 3.01 and Carty 2008 ?
And of course some screenshots from Carty , thanks .
Thanks Carty dloadin now, nice to see a new rom
Spartanec said:
Yeah... We want screenshots! And info about memory. What about Office..?
Click to expand...
Click to collapse
I thought of removing Power Point alone which was taking around 3MB and none may use it. But finally left it. So the ROM has office for the professional use! Since most of the things were removed, I thought a screenshot may not be required. Anyway, I'l update the first post with a shot taken now.
Storage Memory: 51.45/56.80 MB Free
Program Memory: 18/50 MB Free
Dont matter about the Program Memory, it is due to the page pool optimized to run a lot faster. Many ROMs start with 26 MB of Free Program Memory but later leaks down to 15 and the device becomes slower. By ROM never went slower after a day of use and was still at the same speed. Maybe, I shouldn't boast. So will await your test responses.
Regards,
Carty..
Thank you I'm downloading. Again Big..Big thank...
Great job !
And for information, OS 5.2.2000 is not from MS but from pdaviet team !
What about voice dial?
How much is the PagePool?
Nice one, back to basics kind of stuff, just what we need!
BTW, did u include SDHC support?
dlding now.
Olegas said:
What about voice dial?
How much is the PagePool?
Click to expand...
Click to collapse
Page Pool is 6MB but I'm not very sure. Voice Dial is removed.
kakainis said:
BTW, did u include SDHC support?
dlding now.
Click to expand...
Click to collapse
I think I included a SDHC dll file but I'm not sure if it works as I don't have an SDHC card. Since this is my first ROM, I didn't include anything extra which I thought is not useful for all. I will watch for what people want here and build them in the next ROM.
Meanwhile, does somebody have a tutorial on XIP and SYS porting? I saw one on the Atom forum but that wasn't complete.
Regards,
Carty..
Mirrors please
Could I ask for another mirrors by Megaupload, thank you.
!
I like your work, thank you
cheers very muchly indeed...
carty did you use meschles' vanilla 2.0/.1 or touch vanilla 3.0/.1, if so, is this faster and/or as stable?
or anyone else know?
thanks again for your time spent on this
Thank you
Thank you very much for releasing this fastest rom on WM6
But i encounter an error on it my 2GB sd card stop working after i upgrade it
can anybody please HELP ME OUT how can i make it work.. PLZZ
Thank you in advance
It stopped working? Maybe its an SDHC card?
Regards,
Carty..
Well, it's about time for the X50v!
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Ok, I know I've been lagging on working on this device here lately. But apparently even though I don't know how this happened, or why it waited so long to happen, I've finally have been able to build you all a working ROM for the X50v. This is a test release. This is not going to be my final A09 release for this device. So please test the crap out of it, and notify me of any bugs/problems. Only thing I think might be wrong with this ROM, is the IrDA. But I'm working on that. But test it anyway.
Enjoy!
Test #1: Click here to download
Test #2: Click here to download
*** Changes in Test #2 ***
1. Turned Off Error Reporting (Forgot to remove Icon in settings)
2. Removed/Turned Off SQM A.K.A. Customer Feedback
3. Added WM6.1 Speed Enhancements (Partial INITVMMAP)
4. Added Wireless Power
*** Test #1 BUGS **
1. Not sure IrDA is working.
2. I forgot to include .NET 2.0. So Some apps do not load. It is attached below.
*** Test #2 BUGS **
1. Not sure IrDA is working.
2. .NET 2.0. Not Included (Build Errors). So Some apps do not load. It is attached below.
Yay! Fantastic news! Thanks a bunch Lennysh! I'll try it, well, asap, which unfortunately will probably not be in the next few days
But again, many thanks for your work, great news for all X50v owners!
Cricri said:
Yay! Fantastic news! Thanks a bunch Lennysh! I'll try it, well, asap, which unfortunately will probably not be in the next few days
But again, many thanks for your work, great news for all X50v owners!
Click to expand...
Click to collapse
Kewl, might be best anyway. I should have a newer one out later on today I hope. Should be even faster than this one...
Is this one UC based?
mcdull said:
Is this one UC based?
Click to expand...
Click to collapse
No, but next release due out today I hope will be. All my X51v ROMs are UC. This one I released cause I was so happy to get it working. But it is a good copy once you install .NET.
lennysh said:
No, but next release due out today I hope will be. All my X51v ROMs are UC. This one I released cause I was so happy to get it working. But it is a good copy once you install .NET.
Click to expand...
Click to collapse
Thanks for your reply... I am NOW updating my device. But I guess it will take forever to load it in to the NOR. The biggest trouble I am having is the slow down and 100% CPU issue with Maglite A03. Hope this one works better.
mcdull said:
Thanks for your reply... I am NOW updating my device. But I guess it will take forever to load it in to the NOR. The biggest trouble I am having is the slow down and 100% CPU issue with Maglite A03. Hope this one works better.
Click to expand...
Click to collapse
Well, Like I said, my next test release will be faster than this one. And I will be researching the FILESYS thing also, so maybe, who knows, I might be able to tune that down a little bit. I do not think the FILESYS problem exists on every NOR device running WM5+.
lennysh said:
Well, Like I said, my next test release will be faster than this one. And I will be researching the FILESYS thing also, so maybe, who knows, I might be able to tune that down a little bit. I do not think the FILESYS problem exists on every NOR device running WM5+.
Click to expand...
Click to collapse
After reading lots of tech document, I now convinced that the filesys / compaction issue is not avoidable. But fine tuning the cache / file system could help a bit. It is just a damn thing that Dell / HP just don't want to change the 4K sector for NOR device.
mcdull said:
After reading lots of tech document, I now convinced that the filesys / compaction issue is not avoidable. But fine tuning the cache / file system could help a bit. It is just a damn thing that Dell / HP just don't want to change the 4K sector for NOR device.
Click to expand...
Click to collapse
Well, I've only recently used the X50v any really. So I don't know much of this problem. But if there is any regedit's that you know of that helps, and anything you think that can be done, please post. It will not hurt me a bit to add what I can to the ROM, and try anything at all to improve this "bug".
lennysh said:
Well, I've only recently used the X50v any really. So I don't know much of this problem. But if there is any regedit's that you know of that helps, and anything you think that can be done, please post. It will not hurt me a bit to add what I can to the ROM, and try anything at all to improve this "bug".
Click to expand...
Click to collapse
Most of the tweaks don't actually work. The issue is related to the time take to erase used block in NOR. It take few seconds for one block and therefore it will not erase until it is actually needed to be used. And when it is out of available block, everything will stop working (100% CPU load) until some free blocks are available. So the only solution is to put as little file as possible into the ROM, and put EVERYTHING changeable items into the SD CARD.
mcdull said:
Most of the tweaks don't actually work. The issue is related to the time take to erase used block in NOR. It take few seconds for one block and therefore it will not erase until it is actually needed to be used. And when it is out of available block, everything will stop working (100% CPU load) until some free blocks are available. So the only solution is to put as little file as possible into the ROM, and put EVERYTHING changeable items into the SD CARD.
Click to expand...
Click to collapse
Post all the regedits needed to set the SD Card/CF Card as the target folder for anything you think we need to. I know you can do this for cache/temporary internet files. I just want to see all the one's that the X50v user's think are necessary to try and help this as much as possible.
Change pagepool to 0M will make filesys happier. Filesys may still use a lot of CPU sometimes, but the response delay during that time will be acceptable. Can you change the pagepool to 0M?
quick post for bug.
Wifi works only when switched off and enable with wifi button.
It cannot be controlled with software only.
Same happens in MagLite ROM
yboy said:
Change pagepool to 0M will make filesys happier. Filesys may still use a lot of CPU sometimes, but the response delay during that time will be acceptable. Can you change the pagepool to 0M?
Click to expand...
Click to collapse
Well, I usually set all my pagepools to 5 megs for X51v, but I will post a pagepool changer with this rom added to the list in it. That way people can change it themselves depending on what they like.
mcdull said:
quick post for bug.
Wifi works only when switched off and enable with wifi button.
It cannot be controlled with software only.
Same happens in MagLite ROM
Click to expand...
Click to collapse
It's using same OEM folder from MagLites Kitchen. So this is why that still is a bug. I will be working on this to try and fix.
That will be great. The slowness is mostly caused by write to the page pool, so change it to 0M really helps. Besides, all changeable file should be relocated to CF / SD card, by registry or other tricks. That will make x50v faster.
x50v is largly different from x51v becoz of the NOR flash.
It is simply unusable when filesys kicks in..
My experience is that.. with 80MB free storage, sometimes I need to spend 5 minutes to start opera and get into google (the first page). Which is simply... unacceptable.
yboy said:
Change pagepool to 0M will make filesys happier. Filesys may still use a lot of CPU sometimes, but the response delay during that time will be acceptable. Can you change the pagepool to 0M?
Click to expand...
Click to collapse
Page Pool Editor from paveld has been attached to first post of this thread with X50v A09 ROM added to the list at the bottom. As you will see, the default Page Pool for this ROM was 4 megs already. So that is what you all are using right now. But least now you can change it, then re-flash. Let me know how this works for you.
yboy said:
That will be great. The slowness is mostly caused by write to the page pool, so change it to 0M really helps. Besides, all changeable file should be relocated to CF / SD card, by registry or other tricks. That will make x50v faster.
Click to expand...
Click to collapse
Can you post those regedit's since you probably know all the best ones to use?
I have disabled cache of IE (Using Tweaks2K2.NET) and NetFront, and install everything on CF card. This helped a lot.
[APP] RogueTools v1.1 - [Overclock & Change Screen Density] - [Updated: 04/27/2010]
I now present you with RogueTools v1.1.
Summary:
This application is designed for the Vogue, Kaiser, Polaris and Nike devices. It allows you to dynamically overclock your device as well as change the screen density.
Requirements:
This application requires write access to your /system so this will only work on EXT2 and NAND installations.
Overclocking Notes:
Overclocking results will vary. Based upon the beta test results everyone was able to Overclock to 500mhz and have very stable results. Above this results really were mixed. Give it a try, I'd love to hear your results.
To download, search in the market for "RogueTools"
Screenshots:
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Before (400mhz) and After (500mhz) CaffeineMark Benchmark
HUGE thanks to all of the beta testers. You guys did an outstanding job!:
citrit, cp0020, damianiw, DmK75, dzo, GriloVirante, Incubus26jc, inertone, jnadke, mblaster, Millence, mssmison, paipo, pyrator, scouzer, vrflyer
Disclaimer:
As always, I’m not responsible for anything that happens to your phone while using these files.
Future Versions:
I'm looking for some creative ideas on other additions you'd like to see added to this application. So if you have some thoughts, please let me know.
Changelog:
04/27/2010
Added device support for Nike
04/24/2010
Version 1
Awesome. Can't wait to try.
Will this affect the battery life?
Right now the battery is bad enough with android, but it does sound really nice.
Nice,
will try it out
Overclocking will always effect battery life negatively. Under-clocking will increase it.
You don't always need your Kaiser running at 400 Mhz, but I think Android throttles it anyway.
Dukenukemx said:
Overclocking will always effect battery life negatively. Under-clocking will increase it.
You don't always need your Kaiser running at 400 Mhz, but I think Android throttles it anyway.
Click to expand...
Click to collapse
When sleeping it underclocks to ~200Mhz (it seems that "cat /proc/cpuinfo|grep BogoMIPS" displays something around the clock value).
nice work bro...works great on my tilt
could you upload the update.tar ?
Tukang PDA said:
could you upload the update.tar ?
Click to expand...
Click to collapse
Can't You read?
To download, search in the market for "RogueTools"!!!!
Awesome.
Kaiser
Benchmark
400MHz : 91 points
475MHz : 130 points
500MHz : 140 points
Nice!
Appears to work on my tilt!
Nice Job!
liveskelet said:
Awesome.
Kaiser
Benchmark
400MHz : 91 points
475MHz : 130 points
500MHz : 140 points
Click to expand...
Click to collapse
600Mhz : Called fire brigade....
zenity said:
600Mhz : Called fire brigade....
Click to expand...
Click to collapse
Are you serious?
I may have to raise the max threshold for you Kaiser users. 700???
Holy crap this works awesome. Tried it on Polymod 2.1d and it kinda blew my mind. I've only overclocked to 450mhz so far, but it's a night and day difference. Thanks a ton myn, great work.
tonyc0642 said:
Holy crap this works awesome. Tried it on Polymod 2.1d and it kinda blew my mind. I've only overclocked to 450mhz so far, but it's a night and day difference. Thanks a ton myn, great work.
Click to expand...
Click to collapse
I've set mine to 500 Mhz, any chanche for overheating (when going higher then 500)?
(I've deleted some .apk packages (voice dialer,...) and together with this tweak it runs really great, almost 40MB free ram whe killing all apps)
Great app. Can someone give a bit more info about the "increase screen dpi" option? When and why would I want to use this feature? What is a "good" value?
robuser007 said:
I've set mine to 500 Mhz, any chanche for overheating (when going higher then 500)?
(I've deleted some .apk packages (voice dialer,...) and together with this tweak it runs really great, almost 40MB free ram whe killing all apps)
Click to expand...
Click to collapse
I can't say for sure since I'm not the one that developed this, but it seems to me that it should actually save your battery since it underclocks when the phone is sleeping, unless of course you're actively using it a lot while at 500mhz or more, in which case you'd use your battery quicker and run the risk of overheating. Like I said, I can't be certain, so you may want to wait for a reply from myn himself on it.
jt131 said:
Great app. Can someone give a bit more info about the "increase screen dpi" option? When and why would I want to use this feature? What is a "good" value?
Click to expand...
Click to collapse
It should be set at 120dpi by default, at least it was for me. The dpi is "dots per inch," which is pretty much what it sounds like. If you make the value higher, it will make things larger. If you make it smaller, well you get the idea. I played around with it some to check it out, but I'm happy with the default dpi.
Thanks tony. I guess it could be useful for applications that do not work correctly in Kaiser's QVGA, like iGO for example. Would a lower dpi setting compensate for iGO's larger resolution design?
Reducing the DPI makes everything small so iGO might work but it will make reading everything harder.
jt131 said:
Thanks tony. I guess it could be useful for applications that do not work correctly in Kaiser's QVGA, like iGO for example. Would a lower dpi setting compensate for iGO's larger resolution design?
Click to expand...
Click to collapse
Who needs to read, so long as you can see the map and hear the voice commands?
This thread has the objective of testing the latest Hero's ROMs available, through CPU benchmarking.
To perform these tests the following applications have been used:
Linpack 1.1.2 - GreeneComputing benchmark application for CPU and Dalvik VM performance. Measured in MFLOPs. [link]
BenchmarkPI 1.11 - Application used for benchmarking CPU performance. The objective is to calculate the 1st n digits of PI. Measured by a timer. [link]
Benchmark 1.03 - Softweg's test suite for benchmarking. All tests, except the GPU one, done here to account for CPU, Memory and File System performance. Measured by an internal score. [link]
All the ROMs were clocked @ 691MHz "Performance Mode" under SetCPU [link], unless specifiied otherwise.
After each ROM installation, the 4 applications were installed, Airplane mode was turned on, the phone was rebooted 2 times, and I clocked the 3rd boot time.
5 dummy tests were made prior to the real 5 valid tests, per application, so the results would be clearer.
The higher the score, the better, except for BenchmarkPI, since it is a timed test.
Non-JIT and non-BFS kernels were used in these tests.
Also, I forgot to time the Official 2.1 Boot time, my bad. Feel free to do it.
This work was made as a complement to Arkymedes 3D/OpenGL benchmarks. See his thread [here].
>> If you have a ROM request, just post it and I will test it. <<
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Changelog
Changelog @ 15/July/2010:
HebVillain 1.1.2 [link]
RoboMIX 0.5.5 [link - android.com.pl]
Changelog @ 13/July/2010:
Chocolate Eclair 1.1r2 [link]
RoboMIX 0.2.1 [link]
Changelog @ 12/July/2010:
Fallah 1.3 [link]
Changelog @ 11/July/2010:
Chocolate Eclair 1.1 [old link]
Changelog @ 9/July/2010:
aHero 1.1 [link]
Androbin CR4 - CMH 5.0.7 [link]
Chocolate Eclair 1.0.1 [link]
Cronos Droid 2.0.2 [link]
Cyanogen Mod 5.0.8 [link]
Daguu Espresso 1.3 [link]
Daguu Sense 1.0 [link]
EzHero 5.0.1 [link]
HeroDroid 0.7 [link]
Kimera 2.1 [link]
Official Android 2.1 [link]
SuperHero 2.0 [link]
Vanilla Eclair 4.0 [link]
VillainROM 12.0.0 [link]
Looking forward to the charts.
Holding off trying new ROMs for your results.
scouzer said:
Looking forward to the charts.
Click to expand...
Click to collapse
quote for subscribe
Graphs added. Enjoy.
respect your work. Very nice work done here. Thanks
Feeyo said:
respect your work. Very nice work done here. Thanks
Click to expand...
Click to collapse
I'm tired, lol.
Almost 12 hours straight (with some pauses, obviously).
Thanks for the comment.
As usual Kimera shows up and kicks the ass off whole dev teams.
Very interesting work. Thank you very much. You must have worked hard.
scouzer said:
As usual Kimera shows up and kicks the ass off whole dev teams.
Click to expand...
Click to collapse
Really?
From looking at those results I thought it was interesting to see there wasn't one clear cut better rom as they all seem to excel in different areas..
Although considering super hero is based off vr12 I was shocked to see sense adding 10 secs to the boot time....
As im running vr12 but using launcher pro and no sense widgets might aswell switch over...
Thanks for putting in the time to do this!
Very nicely done.
Nice work.
Only comment (take it as constructive criticism btw for next time, don't worry bout it), is that in the last graph, it's really pretty hard to see which bars are for each ROM. It may be my eyes, but I just felt it harder than the others. A little gap between each possible?
Nice work btw.
I was also interested to note that VR12 is 10 seconds longer to boot than SuperHero, despite both being based on very similar base ROMs, and made in the same way.
anon2122 said:
Nice work.
Only comment (take it as constructive criticism btw for next time, don't worry bout it), is that in the last graph, it's really pretty hard to see which bars are for each ROM. It may be my eyes, but I just felt it harder than the others. A little gap between each possible?
Nice work btw.
I was also interested to note that VR12 is 10 seconds longer to boot than SuperHero, despite both being based on very similar base ROMs, and made in the same way.
Click to expand...
Click to collapse
I thought SuperHero was a stripped down version of VR10.3?
Great Job! Love it, but could you add Jit to them to?
xenoside said:
Great Job! Love it, but could you add Jit to them to?
Click to expand...
Click to collapse
JIT is unstable and tends to lag the menus performance, though it increases the Linpack score. It may be seen has a ghost improvement, since you pratically don't have any advantage with JIT enabled.
I speak for myself. I prefer an overall fluid ROM than a JIT enabled one, that will FC every 10min and reboot every 30min.
Correct me if I'm wrong.
shokkapic said:
JIT is unstable and tends to lag the menus performance, though it increases the Linpack score. It may be seen has a ghost improvement, since you pratically don't have any advantage with JIT enabled.
I speak for myself. I prefer an overall fluid ROM than a JIT enabled one, that will FC every 10min and reboot every 30min.
Correct me if I'm wrong.
Click to expand...
Click to collapse
I vote for this benchmark to be STICKY! Seriously good job and well worth reading.... Cheers m8!
I'd like to see results with and without JIT enabled.
It would be a good measure of the performance improvements/losses possible on different roms (sense/AOSP etc..), and maybe even throw in a "# of FC's with JIT enabled during benchmarks" graph to show how stable/unstable it is.
Also maybe a game/fraps type benchmark would be useful?
Nothing like tripling all the hard work you're doing
rakig said:
I'd like to see results with and without JIT enabled.
It would be a good measure of the performance improvements/losses possible on different roms (sense/AOSP etc..), and maybe even throw in a "# of FC's with JIT enabled during benchmarks" graph to show how stable/unstable it is.
Also maybe a game/fraps type benchmark would be useful?
Nothing like tripling all the hard work you're doing
Click to expand...
Click to collapse
The same way the results are very similiar without JIT @ Linpack application, the same similarity would appear with JIT enabled.
About the # of FC's, I would need a lot more time than 30 / 40min per ROM (the amount of time I used in these sequence of tests), and that is time I simply don't have, for now. To test the # of FC's and restarts I would need something like 3 hours of intense use per ROM, opening a whole bunch of different apps, basically "wasting time" until something goes wrong.
About the game/fraps type benchmark, Arkymedes is doing something similar. Read again the first post. You will find the link to his post there.
anon2122 said:
Nice work.
Only comment (take it as constructive criticism btw for next time, don't worry bout it), is that in the last graph, it's really pretty hard to see which bars are for each ROM. It may be my eyes, but I just felt it harder than the others. A little gap between each possible?
Nice work btw.
I was also interested to note that VR12 is 10 seconds longer to boot than SuperHero, despite both being based on very similar base ROMs, and made in the same way.
Click to expand...
Click to collapse
I will try to fix the last graph. Since it is made in Google Docs and updated automatically, it doesn't have the amount of editing settings as in Microsoft Excel.
I had to repeat the VR12 boot for a 4th time, because it was taking 90+ seconds, already overclocked @ 691Mhz on boot. The 4th time was the best I got from all the other 3. Maybe if there was a 5th try, the results would be more similar. I will check it now, since VillainROM was the latest rom to test, and I hadn't flash any since yesterday. I will edit this post in 2min, just to boot it once more. BRB.
Edit: 73 Seconds... The 70 seconds boot was made with Airplane Mode on. This boot wasn't.
----------------------------------------------------------------------------
I'm very glad you guys are taking my "little" work so serious. It means a lot to me.
Also, the last graph was updated, and the differences between the ROMs are clearer now.
Thank you all, once again.
Great stuff, the boot times made me test Vanilla Eclair 4 and it rocks (UI was more responsive to me than ezHero 4.0)! I'd suggest to add (for the future) the amount of free RAM after rebooting with standardized autokill settings. Vanilla Eclair scored very very well here, better than other rom I have tested.. a shame I cannot get the HTC dialer in vanilla roms :-(
Awesome work I'm happy to see daguu roms score that high - maybe this will help flzyup gain some popularity because apparently his work is on par and sometimes better than most of other devs' work.