[Q][Resolved]1 core / 2 core...? - Atrix 4G Q&A, Help & Troubleshooting

I am currently running the latest stable cm7 (Ba2tF), I found a program (quadrant) that gives me statistics about my phone.
What concerns me is it says I have 1 core. Is this the phone not utilizing resources? If so, is there a known method to force it to use it? Or is it possible the app doesn't register the second core for some reason?
Or is this something I messed up by being a relative n00b??
{
"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"
}
Sent from my MB860 using xda premium

jeranon said:
I am currently running the latest stable cm7 (Ba2tF), I found a program (quadrant) that gives me statistics about my phone.
What concerns me is it says I have 1 core. Is this the phone not utilizing resources? If so, is there a known method to force it to use it? Or is it possible the app doesn't register the second core for some reason?
Or is this something I messed up by being a relative n00b??View attachment 772122
Sent from my MB860 using xda premium
Click to expand...
Click to collapse
quadrant does not work well with dual core devices. use apps like cf-bench or antutu

Cheers! I'll check them out.
Sent from my MB860 using xda premium

Okay, let me make this clear, because it's not written in any FAQs I've seen, and it took me a while to work out: Android (at least on Tegra 2 devices) DISABLES CPU cores when they are not needed. Garbarge collection has been moved exclusively to core 0 so that core 1 can remain idle for extended periods and be disabled when not needed. Owing to the fact that looking up system information is not a particularly strenuous, multi-threaded task you will almost NEVER see the second core listed unless you've only just come from another task which was using loads of power and it hasn't yet got round to 'un-latching' the second core.
Look in your system log, you'll see it's full of messages about Core 1 being attatched/detatched.
My only issue is not understanding this behaviour but why the decision to implement it was made. I'm fairly certain it must have overheads and I'm also pretty sure NVidia stated that the CPU cores in Tegra 2 are not individually power gated - i.e. there should be no power saving benefit to doing this...?

Azurael said:
Okay, let me make this clear, because it's not written in any FAQs I've seen, and it took me a while to work out: Android (at least on Tegra 2 devices) DISABLES CPU cores when they are not needed. Garbarge collection has been moved exclusively to core 0 so that core 1 can remain idle for extended periods and be disabled when not needed. Owing to the fact that looking up system information is not a particularly strenuous, multi-threaded task you will almost NEVER see the second core listed unless you've only just come from another task which was using loads of power and it hasn't yet got round to 'un-latching' the second core.
Look in your system log, you'll see it's full of messages about Core 1 being attatched/detatched.
My only issue is not understanding this behaviour but why the decision to implement it was made. I'm fairly certain it must have overheads and I'm also pretty sure NVidia stated that the CPU cores in Tegra 2 are not individually power gated - i.e. there should be no power saving benefit to doing this...?
Click to expand...
Click to collapse
I reckon not having a second core running at 300-800mhz just to use the calculator/calendar would massively outweigh the "overheads" involved in enabling/disabling the 2nd core. I'm not sure about the individual power gates though

Related

Memory bandwidth available to ARM Cortex-A9 cores in Exynos 4210?

Supposedly it should have dual channel 32-bit 800Mbps (the same as OMAP4). But apparently utilizing full memory bandwidth might be a bit challenging with ARM Cortex-A9 (at least early revisions). One example is OMAP4: groups.google.com/group/pandaboard/msg/bd03264b6b800900 And another example is here: anandtech.com/show/4225/the-ipad-2-review/4 (just look at the "Stdlib Write" vs. "Stdlib Copy" numbers, somehow "Stdlib Copy" is unexpectedly low for newer iPad 2).
So could anybody from early adopters benchmark memory bandwidth in Exynos based Galaxy S2? It would be very interesting to see the results for memcpy/memset ("Stdlib Copy"/"Stdlib Write"). There are some benchmark applications in Android Market, and one of them which is called "Benchmark" (softweg) seems to be able to provide "Copy memory" statistics.
I'm going to make a decision whether to buy this phone based on this information
Thanks.
{
"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"
}
What did the ipad get?
ice_coffee said:
What did the ipad get?
Click to expand...
Click to collapse
It's difficult to compare results of different tests even if they attempt to measure the same thing. But the whole point is that "Stdlib Copy (single-threaded scalar)" benchmark regressed from 740.6 MB/s to 522.0 MB/s in iPad2 when compared to the original iPad according to the data from Anand's article that I linked.
According to the data provided by Intratech (thanks a lot!) Galaxy S2 shows throughput ~700MB/s for memory copy performance in this particular 'Benchmark' program from the Android Market. And on my Galaxy Tab (Hummingbird, Cortex-A8) the same test program shows ~560MB/s. Looks like Exynos is at least somewhat better than Hummingbird, which is very encouraging, considering the troubles encountered by the other vendors. Here is one more link for OMAP4 - computerarch.com/log/2011/03/01/pandaboard/
And this particular benchmark program is not a perfect one. On Galaxy Tab, I can get better memory bandwidth for memory copy by using NEON instructions and boost it to something like ~670MB/s (up from ~560MB/s shown by the benchmark). Which additionally shows that the results from different tests may vary a bit and it's better to run the same test program when comparing different hardware. Also memory latency for random memory accesses is quite important and I don't know any existing test in Android Marked which could benchmark it, but I hope that Samsung also managed this right
With all that said, such low level benchmarks cover only a single aspect of system performance. And good memory bandwidth is important only for certain types of applications. I was just curious whether Exynos could possibly have any memory bandwidth issues or not.

[Q] Can we get higher overclocking frequencies?

Can our phone go higher than 864 mhz?because in the lg optimus gt540 there support for 900 mhz and 1000 mhz.
yoman321 said:
Can our phone go higher than 864 mhz?because in the lg optimus gt540 there support for 900 mhz and 1000 mhz.
Click to expand...
Click to collapse
Absolutely. I think 2GHz is easily feasible if you plug the phone to 220V directly instead of using the wall charger!
http://forum.xda-developers.com/showthread.php?t=1182804
Check out this rom for that phone
{
"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"
}
@Dear MIA moderator, kindly lock this thread if a miracle ever happens and you appear.
I have read somewhere that higher frequencies can be faked as well. So some roms might be behaving like this.
On the other hand, I have read 1gh oced to 1.8gh. An 80% OC.
So, theoretically, our phone oc can be stretched to 600*1.8= 1080mh.
Depends on many other factors as well, which, I don't know.
Sent from my LG-P500 using Tapatalk
Its not faked here's the video
http://www.youtube.com/watch?v=GSbCIvp5fzA
yoman321 said:
Its not faked here's the video
http://www.youtube.com/watch?v=GSbCIvp5fzA
Click to expand...
Click to collapse
We've so many posts complaining about kernel panics for anything more than 748 MHz(even 710 in some cases). If the phone doesn't support it, accept the fact and move on. This is not a super phone and has its limitations. So practically, it does not make sense to do something the phone is not capable of, unless you want to have your device getting bricked forever.
yoman321 said:
Its not faked here's the video
http://www.youtube.com/watch?v=GSbCIvp5fzA
Click to expand...
Click to collapse
Listen, yup optimus gt540 can achieve to 900 mhz wit a same processor BUT ours are in a different hardware... Like gt540 has a imageon aderno 200 gpu series like optimus one is on qualcomm aderno 200 series..
Sent from LGE Secret phone using tapatalk app...
yoman321 said:
Its not faked here's the video
http://www.youtube.com/watch?v=GSbCIvp5fzA
Click to expand...
Click to collapse
Come back after you have found a P500 phone that is stable at at least 864MHz. Good luck. After you have tested couple hundreds of phones with no success, perhaps you will go find yourself another ricer project.
Can't even see the practical purpose of getting our cpu at these high freqs.
This phone has a limited hardware capabilities compared to high end devices. Everything that this phone can handle, can do it pretty well with frequencies way less than 800mhz (combined with a nice kernel/rom setup ofc).
Personally i'm one of the few ones with a phone stable at 864 mhz. But for anything except gaming emulators or mkv movies decoding there is no real uses in what i do to maintain it at that speed. The changing of screens on the launcher is also a little bit faster at 864 and thats it. For 150$ and wifi tethering its still amazing.
Sent from my LG-P500 using XDA App

CPU 1.3 or 1.5 GHz?

In the specs it says: CPU runs with 1.5 GHz. But my phone runs with 1.3 GHz... :crying:
Bug? Wrong specs? What is your CPU running at? 1.3 or 1.5 GHz
Besides: The phone is fast as lightning. Just want to know
From what I heard it's real MT6753 1.5 GHz but downclock at 1.3 GHz due to battery power. If this is true, I can say very, very stupid reason.
{
"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"
}
xdazilla said:
In the specs it says: CPU runs with 1.5 GHz. But my phone runs with 1.3 GHz... :crying:
Bug? Wrong specs? What is your CPU running at? 1.3 or 1.5 GHz
Besides: The phone is fast as lightning. Just want to know
Click to expand...
Click to collapse
well, 1.3 is the specs I have seen and that is the same as the phone I have.
frankly I don't need any more speed at the moment... as this phone poops on phones for similar money available over here in the UK.
You could get the ZP999 for even more 'Zopo speed' but then you'd have an older OS, and a non removable battery... and lots of people complaining about having the battery run flat in a few hours... I think the Speed seven is the one to have!
Real specs..
Enviado do meu ZP951 através de Tapatalk
Netmatrix said:
Real specs..
Click to expand...
Click to collapse
This still doesn't add up... Droid Info app shows there are 8 processor cores, and I can see it is possible to use all 8... but MT6735 only has 4 cores. How do you explain this? Does the phone have two MT6735 chips?
Someone from ZOPO should explain this.... it's an 8 core or a 4 core phone really?
If you look at the screenshot, it mentions first MT6753 and the reports the CPU as MT6735. So there is certainly something wrong and we can not trust it.
Just check the phone with AnTuTu a Geekbench and compare the results with other MT6753 phones. That would make it clear.
Enviado do meu ZP951 através de Tapatalk
In my Zopo Speed 7 plus
For the moment I can not send image (rule of the first 10 messages).
But mè cpu-z sees as 1.30 GHz:crying:
The revision of the processor is the same r0p4
@Netmatrix
Or you tinkering with the new kernel, and / or source code released?
it's my modified kernel but i don't have any speed improvement, and lowered the lowest stage at 199 Mhz
Netmatrix said:
it's my modified kernel but i don't have any speed improvement, and lowered the lowest stage at 199 Mhz
Click to expand...
Click to collapse
Maybe it helps the battery to last longer!?

Сheck SPEED BIN (aka PVS) with your OnePlus 2

Some users say that its a smartphone is not heated, others say it is very hot.
It depends on what you got 810 processor.
you need:
1) install AK kernel
2) install Synapse
3) reboot!!!
4) check SPEED BIN in Synapse
{
"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"
}
What is Speedbin/PVS? Can you explain more? I'm also using Synapse with AK.
I have A53 @ 0 and A57 @ 1.
Lower number means that more voltage is needed to certain hertz, and bigger number vice versa. So bigger number is cooler cpu.
Sent from my ONE A2003 using Tapatalk
I am not sure this is as accurate as the prior Speed Bins on other processors. I have had two OP2 phones and both had those same numbers as well. It makes sense that the higher power cores are higher quality, but it seems that all devices I have seen were binned like this.
neatojones said:
I am not sure this is as accurate as the prior Speed Bins on other processors. I have had two OP2 phones and both had those same numbers as well. It makes sense that the higher power cores are higher quality, but it seems that all devices I have seen were binned like this.
Click to expand...
Click to collapse
Yes, It looks like that.
I still didn't understand!
the numbers reported as a53 & a57 pvs are generic and originate from the sys/module/clock_cpu_8994/parameters folder, in files a53speedbin with bit value = 0 and file a57speedbin with bit value = 1. im looking into other locations... if someone could run terminal emulator as SU and execute the command " dmesg | grep PVS " without quotes and busybox must be installed
---------- Post added at 07:01 PM ---------- Previous post was at 06:57 PM ----------
nirmalkurienmathews said:
I still didn't understand!
Click to expand...
Click to collapse
the higher the number the higher the chips theoretical overclock potential and lower transistor voltage leak which translates into real world performance increases because the kernel has various voltage tables for the various quality levels of cpu's... thats about the gist of it... not an expert and this is by no means qualcomm official docs... just trying to shine some light and help the best i can
DO NOT PAY ATTENTION TO THE PVS BIN. why you ask? why should i listen to this novice? well, PVS binning is only for custom made qualcomm CPUS, qualcomm normally uses a CPU design for 2 product cycle, eg 800, 801. the 801 was just the higger binned 800 which allowed them to be over clocked, from then they were binned again. stock arm cpus are built by samsung or TSMC, they are not binned and built with lower tolerances. custom cpus design by qualcomm have slightly higher tolerances. everyone will have a bin 0 a53 and bin 1 A57, purely because they need to be differentiated when made
So nothing special
I've read about PVS on exynos before. Previously it was correct that better values does mean it can tolerate better undervolts and overclocks but on Snapdragon, I'm not sure.
Sent from my ONE A2003 using Tapatalk

Honor 7 graphics GPU problem?

Hi! Suspecting a graphics problem.
So... I just bought this phone. I charged it and updated to 6.0, then installed my apps.
It seemd a bit sluggish when handling graphics, which made me test it with 3dmark. I ran SlingShot ES3.1 and got quite poor results in the graphics section. Physics were fine (CPU related, I guess). Screenshot below. The graphics score was 323 (in one review, which I cant link) and the reference score in app is 283. Mine is 71. I had nothing running background and the phone was recently restarted. What could be the reason?
{
"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"
}
Edit. Power plan is also set to "performance".
As you may know, this SOC does perform well in multitasking but not so in single task and graphics. It's common to all Kirin SOC.
That said, I got 69 score and 1320 Physics, so pretty standard.
It's not a gamer GPU.
zinko_pt said:
As you may know, this SOC does perform well in multitasking but not so in single task and graphics. It's common to all Kirin SOC.
That said, I got 69 score and 1320 Physics, so pretty standard.
It's not a gamer GPU.
Click to expand...
Click to collapse
Actually, good at single and multi core but not on graphics
memht said:
Actually, good at single and multi core but not on graphics
Click to expand...
Click to collapse
Comparatively to Qualcomm apples-to-apples you're wrong, due to it's higher frequency. Check P9 (Kirin 955) and Honor8 reviews when compared with SD820 and SD652.
zinko_pt said:
As you may know, this SOC does perform well in multitasking but not so in single task and graphics. It's common to all Kirin SOC.
That said, I got 69 score and 1320 Physics, so pretty standard.
It's not a gamer GPU.
Click to expand...
Click to collapse
Then how is it possible that the same device has got very different scores when tested in reviews?
futuremark.com/hardware/mobile/Huawei+Honor+7/review
qtba said:
Then how is it possible that the same device has got very different scores when tested in reviews?
futuremark.com/hardware/mobile/Huawei+Honor+7/review
Click to expand...
Click to collapse
Different test settings. Apps installed, accounts, phone settings...
Sent from my PLK-L01 using Tapatalk
But the results surely should not be that bad? I mean, the graphics score I got was less than fourth of what I saw in different reviews. Only difference was, that the tests in reviews were ran with android 5.x still on board and I ran the test with 6.0 just installed. The futuremark test is standard.
My phone was mint condition. No programs on background. Nothing but the standard applicatins installed. Did not make a difference. I ran the test several times just to make sure (with reboots etc.). Socres were 66-71 (graphic score around 53-55).
qtba said:
But the results surely should not be that bad? I mean, the graphics score I got was less than fourth of what I saw in different reviews. Only difference was, that the tests in reviews were ran with android 5.x still on board and I ran the test with 6.0 just installed. The futuremark test is standard.
My phone was mint condition. No programs on background. Nothing but the standard applicatins installed. Did not make a difference. I ran the test several times just to make sure (with reboots etc.). Socres were 66-71 (graphic score around 53-55).
Click to expand...
Click to collapse
Could even be different app versions. Unless you provide the reviews we'll be around endless speculation.
Having different kernel (5.1 vs 6.0.1) also makes a difference.
qtba said:
Hi! Suspecting a graphics problem.
So... I just bought this phone. I charged it and updated to 6.0, then installed my apps.
It seemd a bit sluggish when handling graphics, which made me test it with 3dmark. I ran SlingShot ES3.1 and got quite poor results in the graphics section. Physics were fine (CPU related, I guess). Screenshot below. The graphics score was 323 (in one review, which I cant link) and the reference score in app is 283. Mine is 71. I had nothing running background and the phone was recently restarted. What could be the reason?
Edit. Power plan is also set to "performance".
Click to expand...
Click to collapse
Kirin processors aren't powerful enough for heavy games or single core demanding tasks
PalakMi said:
Kirin processors aren't powerful enough for heavy games or single core demanding tasks
Click to expand...
Click to collapse
True, it´s widely documented throughout the web. Even Mali GPU in latest Exynos are more powerful.
zinko_pt said:
True, it´s widely documented throughout the web. Even Mali GPU in latest Exynos are more powerful.
Click to expand...
Click to collapse
Exactly

Categories

Resources