• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root 1.6Ghz! Its real

See if you can use this info.. On wicked beav's 1600, at 1600, I have weird/weak vibration notifications. Try it out. Lock setcpu in performance and see if the vibration is reduced
Yes I have no noticed random and sporadic issue with haptic feedback fluctuations. More so the back button than the home button. Its intermittent though. Honestly thats not a great bother to me.

Slight tangent here, I dont think I like BLN it seems to me to be a potential battery drainer. Too early to pass final judgement but I might end up getting rid of BLN if I can. Being built into all these kernels not sure thats a choice yet. Im sure someone like wasupwithuman or jt can or already has a disable BLN.
 
OK folks this was intended to be a comparison post between wickedbeav's 1600 kernel and the one's AgentHeX put up earlier. However, I will tell you right here and now the numbers of the so called 2Ghz kernel are worse than his 1600 so Im just going to say at this point for the so called 2Ghz kernel it aint 2Ghz
Now I was prepared to post new pics [ShootMe] I took of the Hex 1600 benchs but I think it will be easier to just make a little chart. I think everyone here will trust my numbers:

Test beav test2 1600 vs. HeX 1600
Linpack 13.6 vs. 11.9
SpeedPi 369ms vs. 410ms
BenchmarkPi 1540ms vs. 1759ms
NeoCore 75.8 vs. 100
NenaMark 55? vs. 89.5
SetCPU Short Bench 109ms vs. 131ms

Quadrant 2557 vs. 2764
CPU 1239 vs. 1166
Memory 1980 vs. 3330
I/O 7746 vs. 7384
2D 487 vs. 479
3D 1333 vs. 1463

OK so conclusions here. As I am on record numerous times before this beav's 1.6kernel scales perfectly with everything we've seen before when comparing it to the hummingbird at 400, 800, 1000 and 1200. I still believe its a 1600Mhz cpu. The benchmarks which are predominantly cpu reliant are Linpack, the 2 Pi ones, SetCPU and of course the cpu portion of Quadrant. In all cases beav's 1600 wins by a decent amount. Also the scaling of beav's 1600's is about a 90% improvement from 800Mhz which is exactly what we'd expect.
Now HeX's kernel does win a few, better 3D for sure, NeoCore, NenaMark and Quadrant 3D clearly show this. So why did HeX win these parts and lose every cpu based test? I believe it has to do with the FSB. I think thats what its called at least for the phone. Obviously we call it FSB for PC's. The higher FSB of the Hex kernal would explain the vastly higher memory scores and potentially the 3D/gpu stuff as well. It could also be the gpu is OC'd on one or both kernels I dont actually know that.
So which kernel is better for daily use? I honestly cannot say at this point. I benched the devil out of the 1600 for almost an hour, I tried the 2000 for 10 minutes and I realized it was worse than the 1600. It seems for 3D games clearly the Hex has an edge. If you are more cpu focused I'd stick with beav. I am going to run Hex now for awhile to grade battery and lag. My initial impression is that Hex has more lag than beav. But dont quote me on that just yet. It really seems like the Hex kernel is closer to a 1.4 given our known scaling. However, if he altered the FSB ratio significantly then many things go out the window.

And having said all that, since the numbers are close enough to each other I wouldn't fault anyone for picking either at this point. Again Im going to test Hex now for a day and see about battery life. That might very well be the difference maker. Both are strong kernel choices at this point.


**update, just set Hex's 1600 to 800 and the numbers scale more closely to beav's than his own. Linpack 6.6, speedpi 770ms, benchmarkpi 3220, quadrant cpu 628. These 800Mhz numbers are much closer to half of beav's 1600 than half of Hex's 1600. Part of me thinks who cares, so HeX has a 1.4kernel or whatever with supercharged FSB and superior 3d gaming.
OK Im out for now.
 
im back on beavs oced version of jts 10/31 voodoo atm. having played with all the variatiosn of the 1600, i still dont know what to think. :confused:

i did notice that the test5 voodoo was in regular use the least laggy and performed the fastest for me, but id have to think the jump from 800 to 1600 (or 1400, whatever it really is) is just too much of a leap and the throttling is off.

Im not sure why the jump to 1600 is even really neccessary. im all about incremental jumps in speed so we the devs, testers, and everyone else can compare more stable and closer speeds. I still really just want to see a 1300 version of jts 10/31 tbh.

fwiw, agenthex is doing a great job and i believe once the bugs/issues are toyed around with by him, jt, etc a lot of what is in his kernels can probably be implemented in the regular releases. i wish i was more knowledgable and could in someway help with that, but alas we all cant be as talented as those guys :rolleyes:

in any event, its great to see all the development/mods going on.
 
Well according to AgentHex the kernel was not really overclocked at all, he was looking into the lag issue and figure it out :(

something to do with slowing down the system and thats why the tests were all high or something (i didnt fully understand it)

while not the same issue as the team whiskey guys, its in the same ballpark id think
 
Now does the same go for beav too? Didnt he just re-compile HeX's?
I dont know the smoothness of beav's is there.
I too ditched HeX's, I was getting that whole battery not connected issue, which we know is kernel related
 
Now does the same go for beav too? Didnt he just re-compile HeX's?
I dont know the smoothness of beav's is there.
I too ditched HeX's, I was getting that whole battery not connected issue, which we know is kernel related

beav's were just recompiled hex kernels so at best they were maybe 1200 voodoo from what i understand.

hex said hes back to the drawing board.
 
beav's were just recompiled hex kernels so at best they were maybe 1200 voodoo from what i understand.

hex said hes back to the drawing board.
Because of the overvolt at top speed I think beav [and hex] uses more battery when you push it. However, their respective undervolt at 100Mhz [phone sleeping] seems to really sip the battery. So Im torn between sticking with this bev kernel or bouncing back to dirrk's. The good news is I wont have to wait long for the next revision.
 
Because of the overvolt at top speed I think beav [and hex] uses more battery when you push it. However, their respective undervolt at 100Mhz [phone sleeping] seems to really sip the battery. So Im torn between sticking with this bev kernel or bouncing back to dirrk's. The good news is I wont have to wait long for the next revision.

Have you tested hex's 1620 ext4 kernel yet?
 
Because of the overvolt at top speed I think beav [and hex] uses more battery when you push it. However, their respective undervolt at 100Mhz [phone sleeping] seems to really sip the battery. So Im torn between sticking with this bev kernel or bouncing back to dirrk's. The good news is I wont have to wait long for the next revision.

yeah hex is playing around with stuff...im currently running jts nonvoodoo 11/2 and its running great. im avoiding voodoo as im playing with stuff, but beav said he *might* make an oc version of the 11/2 if he gets around to it today or tomorrow. til then ill prolly hang with this or flash the voodoo later tonight.
 
Have you tested hex's 1620 ext4 kernel yet?
Nope I have not been on IRC for a few hours so I probably missed 4-5 ROMS and 8-9 kernels by now :p I assume this came down in the recent hew hours? Is this supposed to be the real 1.6 now?
yeah hex is playing around with stuff...im currently running jts nonvoodoo 11/2 and its running great. im avoiding voodoo as im playing with stuff, but beav said he *might* make an oc version of the 11/2 if he gets around to it today or tomorrow. til then ill prolly hang with this or flash the voodoo later tonight.
Thats good news.
 
Nope I have not been on IRC for a few hours so I probably missed 4-5 ROMS and 8-9 kernels by now :p I assume this came down in the recent hew hours? Is this supposed to be the real 1.6 now?

Thats good news.

wait.. there's a fake 1.6?? so the test2 file you uploaded last night, isn't actually 1.6??
 
wait.. there's a fake 1.6?? so the test2 file you uploaded last night, isn't actually 1.6??
Thats apparently what they are saying now despite every benchmark you throw at it affirming it is. At this point I dont get caught up in the banter. Im using either test5 now because even though it locks in NeoCore it does everything else very well. It has the best battery life when not in use, although being overvolted at top speed it does suck it down a lot. Quite frankly there are so many good kernels out now. Shoot I havent even tried jt's 11/2 kernels.

Too many kernels, too little time:D
 
Thats apparently what they are saying now despite every benchmark you throw at it affirming it is. At this point I dont get caught up in the banter. Im using either test5 now because even though it locks in NeoCore it does everything else very well. It has the best battery life when not in use, although being overvolted at top speed it does suck it down a lot. Quite frankly there are so many good kernels out now. Shoot I havent even tried jt's 11/2 kernels.

Too many kernels, too little time:D
so the main difference between test2 and test5 is that test5 locks up in NeoCore? does test5 have better battery life compared to test2?

as for jt's 11/2 kernels, are those @ 1.6Ghz??
 
so the main difference between test2 and test5 is that test5 locks up in NeoCore? does test5 have better battery life compared to test2?

as for jt's 11/2 kernels, are those @ 1.6Ghz??

jts 11/2 are currently standard speed. wickedbeav said he may oc them later on.
 
so the main difference between test2 and test5 is that test5 locks up in NeoCore? does test5 have better battery life compared to test2?

as for jt's 11/2 kernels, are those @ 1.6Ghz??
jt's 11/2 kernels in ROM manager are just 1Ghz, I think slightly undervolted. With respect to wickedbeav's test kernels, when running at top speed [whatever it is] the battery goes a little faster than some other kernels I've tried. However, when it idles its very good. So it will vary on usage. They're not bad at all IMO though. I did notice my clock was a couple minutes off last night, I thought that was weird and might be related somehow to all these issues, I dont know for sure:confused:
 
jt's 11/2 kernels in ROM manager are just 1Ghz, I think slightly undervolted. With respect to wickedbeav's test kernels, when running at top speed [whatever it is] the battery goes a little faster than some other kernels I've tried. However, when it idles its very good. So it will vary on usage. They're not bad at all IMO though. I did notice my clock was a couple minutes off last night, I thought that was weird and might be related somehow to all these issues, I dont know for sure:confused:
so is there any noticable difference between test2 and test5 kernels? other than the locking up of neocore?
 
so is there any noticable difference between test2 and test5 kernels? other than the locking up of neocore?
Not really, maybe, maybe battery life is slightly better with test5. I think that was the only difference between the two is that test5 had tighter/lower voltages
 
Not really, maybe, maybe battery life is slightly better with test5. I think that was the only difference between the two is that test5 had tighter/lower voltages

ah ok, in that case, i'll stick with test2 then.

is it safe to set profiles in setcpu with these kernels? i remember settings profiles before was a no no
 
ah ok, in that case, i'll stick with test2 then.

is it safe to set profiles in setcpu with these kernels? i remember settings profiles before was a no no
You're on your own there my friend. I usually just leave it on conservative unless Im benching it then its performance. Have not messed with profiles because of the same warning.

Also as an add on comment, test5 given its lower voltage [I think] seems to run a bit cooler compared with test2. I'm addict to GRave Defense Gold these days [tower defense type game]. Couple nights ago I was playing like 2hrs of it with test2 and the phone got warm and I was like 100
 
I currently have D02 Kernel LV1000 installed... can i just install this and have no problems??
And where can I find this?
If you are asking if you can flash these alledged 1.6 kernels over your current one yes they're all voodoo.
Find them here:
Test2:MEGAUPLOAD - The leading online storage and file delivery service
Test5:MEGAUPLOAD - The leading online storage and file delivery service
Now unofficially these have been pulled by wickedbeav as there is such controversy over these actually being 1.6. See them for yourself though
 
I'm guessing there aren't any non-voodoo 1.6ghz kernels? I not too long ago completely got rid of voodoo because it was just causing me far too many problems, so I'm not planning on putting it back on.
 
I'm guessing there aren't any non-voodoo 1.6ghz kernels? I not too long ago completely got rid of voodoo because it was just causing me far too many problems, so I'm not planning on putting it back on.

there are nonvoodoo versions...however jts 11/2 nonoc is rom manager is more than capable.
 
I'm guessing there aren't any non-voodoo 1.6ghz kernels? I not too long ago completely got rid of voodoo because it was just causing me far too many problems, so I'm not planning on putting it back on.
There are and if I had any I'd upload them for you; sadly I only grabbed the voodoo ones since Im one of the few who likes voodoo.
there are nonvoodoo versions...however jts 11/2 nonoc is rom manager is more than capable.
You know I actually did try this myself just so I could say I did it thru ROM Manager directly. It was as good as anything else. After a while all the kernels kind of seem the same unless you bench them, but as so many have said what does that really show you these days:confused:
 
Back
Top Bottom