Root Uberfast Snap 7.6 w/ turbo released!

[jmxp69]

battery life is fine. Just remember, this setup is designed to fool the benchmark. It doesn't make your phone run 2x faster than normal. That's why you won't see significant improvements in your normal usage activities, and that's also why there's no more battery drain than usual.

I've seen numbers over 3000 with this "quadrant slayer" build, so technically, all of you guys are still running too slow

You're really off the mark with this. Do you think IO is irrelevant in day-to-day use? What do you think the major bottleneck is in most systems in 2d mode? CPU? Graphics? Not so much.

 

[novox77]

You're really off the mark with this. Do you think IO is irrelevant in day-to-day use? What do you think the major bottleneck is in most systems in 2d mode? CPU? Graphics? Not so much.

off the mark, huh? show me any real-world usage of your evo that is actually 2x faster than stock, as your beloved inflated quadrant score seems to imply.

 

[jmxp69]

off the mark, huh? show me any real-world usage of your evo that is actually 2x faster than stock, as your beloved inflated quadrant score seems to imply.

The only measure we have of improved IO right now is Quadrant, and yes, if the system performs IO faster, the result is a faster user experience. If it's not for you, great, but at least don't spread misinformation.

Edit: BTW Nowhere do we claim a 2x performance boost.

 

[novox77]

The only measure we have of improved IO right now is Quadrant, and yes, if the system performs IO faster, the result is a faster user experience. If it's not for you, great, but at least don't spread misinformation.

Edit: BTW Nowhere do we claim a 2x performance boost.

The reason I am not a fan of the quadrant slayer is because the supposed performance boost, the very thing the benchmark is purported to measure, does not correlate with real-world performance except the benchmark itself.

So you admit the only measure of improved performance is quadrant. And what if the quadrant exploit is flawed? That's like me proclaiming that I am God, and since I am God, I am never wrong; therefore I must be God.

I did in fact try the cm6/snap quadrant slayer build a month ago, and got scores in the 2300s, and I ran real-world speed tests based on actual usage and found.... NO DIFFERENCE in performance. Apps didn't install or load faster. Games didn't run faster. The phone didn't turn into a fkin jet.

Those who follow me on these forums know that I do my homework, and I don't spread misinformation. If you claim that a doubled quadrant score doesn't imply a doubled performance increase, then, tell me, what does a score of 2400 really mean? Provide me with some real evidence of day-to-day performance improvements, and I will happily stand corrected.

 

[novox77]

Why must XDA devs do for HTC what HTC should have done for you?

btw, i see that in your sig, you linked to the original proof-of-concept for the FPS unlock in the kernel.

HTC questioned the kernel mod, claiming that the fps hack only fooled benchmarks, but real-world performance was unchanged. This was easily disproven by several videos (including one I made, in my sig) showing a very significant performance increase: removed all random stutters, introduced smoother scrolling and incredible finger tracking.

This is the type of real-world evidence I'd like to see from a quadrant slayer build. To date, the only evidence I see is people posting their long orange bar screenshot. That's like just showing the bouncing dot in FPS2D. If the quadrant number can legitimately be raised to 2400 (and even 3000+), how does that translate into normal phone operation improvement? Where's the beef?

 

[wake69]

The reason I am not a fan of the quadrant slayer is because the supposed performance boost, the very thing the benchmark is purported to measure, does not correlate with real-world performance except the benchmark itself.

So you admit the only measure of improved performance is quadrant. And what if the quadrant exploit is flawed? That's like me proclaiming that I am God, and since I am God, I am never wrong; therefore I must be God.

I did in fact try the cm6/snap quadrant slayer build a month ago, and got scores in the 2300s, and I ran real-world speed tests based on actual usage and found.... NO DIFFERENCE in performance. Apps didn't install or load faster. Games didn't run faster. The phone didn't turn into a fkin jet.

Those who follow me on these forums know that I do my homework, and I don't spread misinformation. If you claim that a doubled quadrant score doesn't imply a doubled performance increase, then, tell me, what does a score of 2400 really mean? Provide me with some real evidence of day-to-day performance improvements, and I will happily stand corrected.

I am running fresh, netaerchy, and underclocked for battery, and dont notice hardly any difference in perormance than when I ran the cm6-snap setup

Ps it seems more people here are more concerned with battery life than all out speed, at least I see more posts on,battery than speed

 

[jmxp69]

The reason I am not a fan of the quadrant slayer is because the supposed performance boost, the very thing the benchmark is purported to measure, does not correlate with real-world performance except the benchmark itself.

So you admit the only measure of improved performance is quadrant. And what if the quadrant exploit is flawed? That's like me proclaiming that I am God, and since I am God, I am never wrong; therefore I must be God.

I did in fact try the cm6/snap quadrant slayer build a month ago, and got scores in the 2300s, and I ran real-world speed tests based on actual usage and found.... NO DIFFERENCE in performance. Apps didn't install or load faster. Games didn't run faster. The phone didn't turn into a fkin jet.

Those who follow me on these forums know that I do my homework, and I don't spread misinformation. If you claim that a doubled quadrant score doesn't imply a doubled performance increase, then, tell me, what does a score of 2400 really mean? Provide me with some real evidence of day-to-day performance improvements, and I will happily stand corrected.

Every benchmark I've ever used included a set of tests including CPU, Memory, Disk, Graphics. SiSoft Sandra is one of the most respected benchmark programs available, and it measures IO.

Raid0, SSD, HD Cache, Write Caching, Hdparm, and countless other tools are designed to improve IO performance.

My claim is simple: Improved IO = Improved Performance. The ONLY tool available ATM to prove IO performance right now is Quadrant. So are you saying your seat-of-the-pants measure is more accurate? What real world proof do you have that your subjective opinion is more accurate than the only objective measure we have available?

Here are some of my stats: 7.6 has been downloaded 2200 times in 2 days. The snap thread has almost 300,000 views. All revisions of snap have been downloaded over 30,000 times.

Yes we experiment, and yes, sometimes we fail. I freely admit that we are looking under every stone for performance improvements. At the end of the day, we are trying to improve performance, and I believe with the FACTS posted above, we are achieving a degree of success.

I'm not interested in an argument. If it's not for you, this is simple: Don't use it. If you are prepared to tell every person who believes their user experience outside of benchmarking has not actually been improved, I invite you to bring it up in the XDA thread.

J

 

[I[E]

I am running fresh, netaerchy, and underclocked for battery, and dont notice hardly any difference in perormance than when I ran the cm6-snap setup

Ps it seems more people here are more concerned with battery life than all out speed, at least I see more posts on,battery than speed

From 12pm Sunday through 10:30pm Monday I plugged my phone in at 1d 10h 30m 49s. with 12% battery. Cell standby 29% and Phone idle 21%.

Running 7.6cfs Turbo at 998mhz

You can have speed and battery

 

[wake69]

From 12pm Sunday through 10:30pm Monday I plugged my phone in at 1d 10h 30m 49s. with 12% battery. Cell standby 29% and Phone idle 21%.

Running 7.6cfs Turbo at 998mhz

You can have speed and battery

I am not arguing that, I prefer battery, but still enjoy speed. I have no issuues but think arguing ofer a rom and a kernel is a moot point since there are sooooo many differing opinions, just look at the sense v non sense issue, no one is wrong

 

[BlackDave]

Well the high-end gameloft games are running better now (Asphalt was a pain to me)

 

[novox77]

I can see this is a religion to you. After all, if a bunch of people all believe in something, it must be true. Who needs evidence? The kernel has 300k downloads!

You still can't provide any proof or quantifiable measure of what 2400 really means. Your claim is simple:

Improved IO = Improved Performance.

The fact by itself is sound. Who COULD dispute that? What I'm saying is that Quadrant tells us IO is improved. But no one, including you, can quantify improved performance. So, assuming your claim above is valid, doesn't that imply quadrant is wrong? Or more likely, somebody found a neat trick to fool quadrant during its measurement? All I'm looking for here is improved performance. If it were true, I would have stuck with the quadrant slayer build last month.

As you say, sometimes you succeed, sometimes you fail. The kernel uncap was a success. Quadrant slaying, not so much. I'm sorry you can't see that. So, I'm done with this thread and out of your hair, as you requested.

 

[I[E]

Well the high-end gameloft games are running better now (Asphalt was a pain to me)

Impossible! novox77 says it's all smoke and mirrors trickery.

 

[jmxp69]

I can see this is a religion to you. After all, if a bunch of people all believe in something, it must be true. Who needs evidence? The kernel has 300k downloads!

You still can't provide any proof or quantifiable measure of what 2400 really means. Your claim is simple:

I've provided the only measurable or quantifiable outcome in the discussion unlike you who resort to speculation and supposition. You cannot prove IO measurement is wrong can you?

The fact by itself is sound. Who COULD dispute that? What I'm saying is that Quadrant tells us IO is improved. But no one, including you, can quantify improved performance.

Then why bother with ANY benchmark at all? What about linpack, or fps2d. Can you tell me you see the difference between 40fps and 50? Yet we still strive for 60.

So, assuming your claim above is valid, doesn't that imply quadrant is wrong? Or more likely, somebody found a neat trick to fool quadrant during its measurement? All I'm looking for here is improved performance. If it were true, I would have stuck with the quadrant slayer build last month.

This makes absolutely no sense. It's not a trick. It's a measured IO improvement. So you don't accept measured proof, you don't accept perceived proof, what do you accept?

As you say, sometimes you succeed, sometimes you fail. The kernel uncap was a success. Quadrant slaying, not so much. I'm sorry you can't see that. So, I'm done with this thread and out of your hair, as you requested.

Great. Adios.

 

[wake69]

Impossible! novox77 says it's all smoke and mirrors trickery.

If I am correct, what he is talking about is the sd card partition. Overclock inproves processor and kernels can help fps caps. Where it seems that snap picks up most if its score is in read write. Now since it requires a partition, it has the ability to read and write faster. Benchmark tests I have ran with diff rom/kernel combinations seem to be roughly in a similar range. When you add the faster read write this boosts its score. Since all evos have almost identical hardware builds, all will function similarly. That's how I see his argument.

 

[jmxp69]

Now since it requires a partition, it has the ability to read and write faster.

Wrong. 7.6 does not use the sdcard after activation.

 

[wake69]

Wrong. 7.6 does not use the sdcard after activation.

Then why does it require a partition, do u delete the partition after flashing it? Help me understand. My fresh with netarchy hangs on read/write, other than that, it runs well in the test

 

[wake69]

Wrong. 7.6 does not use the sdcard after activation.

Help me understand, I may be misinformed, what am I not understanding?
Pointing out some one us wrong is such a great way to help people understand, it is such a better method to explaining and helping. Thanks in advance for explaining to me where I a wrong

 

[Hazy]

The fact is cm6-snap is faster than any other sense based rom.... 7.6 does not make the phone run 2x as fast as say fresh 3.2 with netarchys kernel, but that's because most of the improvement comes from IO read/write. There's many other components that would have to perform just as fast, like 2d and 3d graphics for there to be and actual doubling of system speed (graphics do perform better with snap, on quadrant or otherwise)

 

[novox77]

I just finished a PM to jmxp69, where I apologized for the tone of my posts above. Re-reading what I wrote, I can see that I ended up fueling more drama than encourage productive discussion about the issue. So I'd also like to apologize to all readers of this thread as well. The spirit of the developers trying to improve the phone's performance should always be commended, and even if the IO improvement is hard to appreciate outside of quadrant, it is still progress.

If you can ignore the emotion and drama from my posts, I hope you can still take away my basic argument, which is that the dramatic increase in quadrant score does not reflect a proportional increase in observable performance, based on my own testing outside of benchmark results. If it's a case of me running the wrong tests, please share.

And IO aside, I am sure that CM6/snap is a very solid build. In fact, it is my most preferred non-stock combo. But honestly, I would like it just as much if it didn't have the turbo.

 

[jmxp69]

I just finished a PM to jmxp69, where I apologized for the tone of my posts above. Re-reading what I wrote, I can see that I ended up fueling more drama than encourage productive discussion about the issue. So I'd also like to apologize to all readers of this thread as well. The spirit of the developers trying to improve the phone's performance should always be commended, and even if the IO improvement is hard to appreciate outside of quadrant, it is still progress.

If you can ignore the emotion and drama from my posts, I hope you can still take away my basic argument, which is that the dramatic increase in quadrant score does not reflect a proportional increase in observable performance, based on my own testing outside of benchmark results. If it's a case of me running the wrong tests, please share.

And IO aside, I am sure that CM6/snap is a very solid build. In fact, it is my most preferred non-stock combo. But honestly, I would like it just as much if it didn't have the turbo.

1) Turbo is optional. You never have to enable it, and you will still get the best .32 build from the snap line I think I've seen yet in terms of battery, stability, and performance.

2) PM reply sent.

3) I cannot argue there is some subjectivity to "feel". The main point behind my case is that I believe we've improved IO, therefore, I believe the end user experience should be improved as well.

4) I also apologize if I came off harsh, so I'll explain myself: v7.6 (if you review the XDA thread you'll find this to be true) was under development and formal beta testing longer than any other snap release. We engaged in a formalized beta program with 12 testers and 5 detailed test patterns. After a week of testing and countless hours of tuning, the group consensus was/is that v7.6 does represent an improved experience. So yes, I got defensive. We invested a massive amount of time in development of v7.6 and it was quite disheartening to see it dismissed out of hand--especially given the work was done by enthusiasts who were unpaid for their efforts.

J

 

[jmxp69]

Help me understand, I may be misinformed, what am I not understanding?
Pointing out some one us wrong is such a great way to help people understand, it is such a better method to explaining and helping. Thanks in advance for explaining to me where I a wrong

wake69: We chose not to explain exactly what we're doing because what we're doing is not subject to GPL and we're considering additional development. For now, turbo is a gift. You are 100% free to take it or leave it.

 

[novox77]

4) I also apologize if I came off harsh, so I'll explain myself: v7.6 (if you review the XDA thread you'll find this to be true) was under development and formal beta testing longer than any other snap release. We engaged in a formalized beta program with 12 testers and 5 detailed test patterns. After a week of testing and countless hours of tuning, the group consensus was/is that v7.6 does represent an improved experience. So yes, I got defensive. We invested a massive amount of time in development of v7.6 and it was quite disheartening to see it dismissed out of hand--especially given the work was done by enthusiasts who were unpaid for their efforts.
J

The only part of the kernel I had beef with was the quadrant score. I have no doubt that the turbo aspect is only a portion of the kernel, and that if it were disabled, snap would still be a damn fine custom kernel. And I definitely appreciate the QA effort put in. I don't question that snap overall provides a better user experience with or without turbo. It worked great for me when I tried it. I just expected a lot more because it did score so high on quadrant.

As a dev, you certainly have the right to take the stance of "take it or leave it; no one's forcing you to use it." I agree with that. You are doing all this work free of charge for your end users. But I don't think it's unreasonable to caution these end users who are flashing their high quadrant scores around that you can't judge the kernel's overall performance based solely on that benchmark, especially when that benchmark puts so much weight into the IO aspect that you have optimized.

For most people, quadrant is a gauge of overall performance. When the quadrant score goes up, the implication is that overall performance is going up along with it. And that's why you see people asking if battery life is ok. Again, the perception is that if something behind the scenes is really cranking to get that high score, it must have some impact on battery life. My observation is that battery life is unchanged (still great, but unchanged) regardless of the presence of turbo.

In no way was I ever implying that snap was a bad kernel hiding behind a great quadrant score. If anything, I am saying that snap is a great kernel, but be careful not to think it's super-awesome JUST BECAUSE it can score high on quadrant. Even I was lured to the kernel initially because of those screenshots. In the end, I'd like to see people come to their conclusions about the kernel without relying on the benchmark score, the piece about snap I feel is misleading a lot of users.

 

[exBBuser]

I have been a snap user ever since the first Kernel came out,I just gotta say that snap does it for me in every single aspect,I don't see myself flashing another Kernel.But I have yet seen any noticeable improvements other than the Quadrant Scores if I compare it with my previous Kernel snap7.5,with this said I'm not saying that the improvements or performance are not there, is just that I don't see them,battery life is the same(well better now that I have my 3500MAh Battery ),Linkpack is about the same and so is FPS2D.Overall performance feels no different than before my pages and apps are opening and loading at the same speeds.

I'm not trying to add more wood to the fire or anything like it. I already said it but I will say it again I'm a sanp lover,just trying to share my real world usage with some of the members.

I'm just an average Joe when it comes to all this Tech stuff but regardless of my skills and knowledge one should always notice any improvements which in my case are none.

In my every day

 

[mitchellvii]

The reason I am not a fan of the quadrant slayer is because the supposed performance boost, the very thing the benchmark is purported to measure, does not correlate with real-world performance except the benchmark itself.

So you admit the only measure of improved performance is quadrant. And what if the quadrant exploit is flawed? That's like me proclaiming that I am God, and since I am God, I am never wrong; therefore I must be God.

I did in fact try the cm6/snap quadrant slayer build a month ago, and got scores in the 2300s, and I ran real-world speed tests based on actual usage and found.... NO DIFFERENCE in performance. Apps didn't install or load faster. Games didn't run faster. The phone didn't turn into a fkin jet.

Those who follow me on these forums know that I do my homework, and I don't spread misinformation. If you claim that a doubled quadrant score doesn't imply a doubled performance increase, then, tell me, what does a score of 2400 really mean? Provide me with some real evidence of day-to-day performance improvements, and I will happily stand corrected.

Since I am the OP on this thread I'll throw in my 2 cents.

You are both right to extent. Benchmarks can be manipulated, but to say a vastly improved benchmark has no effect on real-world performance isn't correct either.

Face it, froma hardware perspective, and EVO can only run so fast. So, if we are to.squeezed more performance from our phones, we need software "tricks" to make it happen. The other day I flashed back to Fresh 3.2 running Kings Kernel #8. After having run cm6 with snap 7,6 for the few days prior I can tell you without question that the Fresh rom was much slower and less responsive. I used to think that rom was really fast but now it seems slow.

Twice as fast? No. Faster? Yeah lots.

 

[novox77]

The other day I flashed back to Fresh 3.2 running Kings Kernel #8. After having run cm6 with snap 7,6 for the few days prior I can tell you without question that the Fresh rom was much slower and less responsive. I used to think that rom was really fast but now it seems slow.

Your test shows that snap outperforms fresh, which was never in contention (by me anyway). A more accurate test would be to compare snap with turbo on, and snap with turbo off. That is how you can observe and quantify the performance gained due to the IO optimization.