Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

what exactly does wiping the davlik and whatnot erase? are all screen settings and everything the same and apps still there?

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

No, it's just cached data that the OS builds over time.

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

well im running damage control, did a davlik and cache wipe, installed and my fps is averaging 43 -47 and multi touch gets 3, but is there a better multitouch viewing app that multitouch visualizer? keeps force closing after i use 3, i think i got 4 but it closed

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

Awesome work

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

I'm running this over SprintLover's ROM. No issues.

I'm just curious, the FPS is different for different tests
I know 2D is simpler. I get about 53 fps on that
On Gears I get about 44-45.
But on Quandrant and Neocore I still get under 30 FPS.

I guess 3D is harder to hack/fix or is that due to the limited GPU? (as compared to hummingbird processor)

Anyway, I absolutely love this. My phone runs so smoothly now.

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

Id like to ask two questions to go along with the above(not directly related).

1. Which is the most accurate tool for evaluating framerates?

2. Upon completing the fps fix, I havent had any issues while in use, but my phone suddenly takes a very long time to do a cold boot. Is this happening to anyone else?

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

Use dotty, by gerry steele, from the market.

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

kernel seems to working very well on OMJ. 4g did turn on as well which is good! Thanks Netarchy!

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

Neocore you should expect around 27 on any snapdragon (even the nexus one benches about 27 in neocore). The neocore score is very much due to limitations of the gpu. In this particular case, the bottleneck for the gpu in neocore is fill rate.

Both quadrant and neocore are made to stress hardware to its limits. They are not necessarily indicative of performance in everything. As an example, neocore may bench at about 27, but I guarantee you if you were to fire up raging thunder 2, you'd find it runs smooth as butter.

As for gears, you should be able to get in the 50's, but it depends largely on the scaling parameters for your cpu.
Benchmarks are highly sensitive to the speed of the cpu, how quickly the cpu governor scales the cpu speed from idle to full in response to load.
For instance, if you were to install setcpu, and set the governor to performance (essentially forcing the cpu to run at full speed all the time) you would find that your benchmarks are consistently higher.

The defaults in the kernel aim for good performance out of the box, but it is worth noting that any default settings can be overridden.

Some custom roms override the defaults by setting the cpu threshold (aka the % of load the cpu experiences before the governor starts increasing the speed of the cpu to deal with the load). It's also worth pointing out that apps like setcpu and overclock widget can also override the defaults, for better or worse (depending on your settings).

The default threshold in my kernel builds is 70%, meaning that once something is throwing a 70% load to the cpu at its current speed, it starts speeding up to deal with the load.

Setting a higher threshold results in better battery life, but progressively lower performance in trade.

Re: [Kernel]netarchy-toastmod, FPS 'fix' for Epson + Nova, with hdmi out - Version 3.
 

1. The most common ones I like to use to gauge real world peak performance are gears and fps2d. They aren't "true" benchmarks in the sense that they aren't trying to push the hardware to its limits and assign an arbitrary score based on that, but they are a bit more realistic in regards of what framerates to expect from the UI, and what many games can achieve. (I say many because there are some games that are just badly made, like almost anything from gameloft ;D).

2.The first time you boot after installing the general population build may take a bit longer than you're used to, as it will be doing a zipalign check on boot and tweaking dalvik-cache. Boot time after that *should* be pretty snappy.