So by over clocking when a certain benchmark tool is detected is perfectly fine because it isn't running for a while? But a game running for extended periods is not because it would cause all sorts of issues?
Samsung know the purpose of these benchmarks and have set it up so they look better than real world usage which put the competitors at a disadvantage.
But then the benchmark tool may trigger it which isn't Samsung's issue. A benchmark test is over with in seconds normally. A game can be run for hours.
So by over clocking when a certain benchmark tool is detected is perfectly fine because it isn't running for a while? But a game running for extended periods is not because it would cause all sorts of issues?
Samsung know the purpose of these benchmarks and have set it up so they look better than real world usage which put the competitors at a disadvantage.
Nothing to do with a benchmark being detected as far as that statement goes the phone is capable of such gpu performance and it is not limited solely to a benchmark as the statement clearly says.
Your statement would imply that the drivers have been altered purely for such benchmarks, that is not in line with the statement
Anand read like there was a white list. The Samsung statement thus sounds badly worded if so.
But assuming that benchmark did some standard video tests that would have run at full speed normally, was the benchmark correct in a few aspects?
I have read the article and as I said from the outset I just don't see it. I understand where they may have reached this conclusion but I don't blindly follow or believe something because its posted.
I do however think that the performance may not be indicative of real world game performance as the gpu is not capable over prolonged periods but this is a long way from doctoring benchmark results. As samsung state the gpu is capable of such speed.
So by over clocking when a certain benchmark tool is detected is perfectly fine because it isn't running for a while? But a game running for extended periods is not because it would cause all sorts of issues?
Samsung know the purpose of these benchmarks and have set it up so they look better than real world usage which put the competitors at a disadvantage.
Sorry I am now confused is it underclocking for games. Apps which do not drain the battery get full GPU
But then the benchmark tool may trigger it which isn't Samsung's issue. A benchmark test is over with in seconds normally. A game can be run for hours.
Nothing to do with a benchmark being detected as far as that statement goes the phone is capable of such gpu performance and it is not limited solely to a benchmark as the statement clearly says.
Your statement would imply that the drivers have been altered purely for such benchmarks, that is not in line with the statement
Sorry I am now confused is it underclocking for games. Apps which do not drain the battery get full GPU
According to Anandtech the phones GPU runs at a standard 480MHz but when certain benchmarks are detected (a whitelist if you will) it overclocks this to 523MHz. The reason it doesn't run at 523MHz all the time is unknown but it would probably cause thermal issues if used for extended periods.
As mentioned by Anandtect
This seems to be purely an optimization to produce repeatable (and high) results in CPU tests, and deliver the highest possible GPU performance benchmarks.
IMO Samsung should either allow games to also run at 523MHz (which wont happen as samsung said it cannot run that for an extended period of time) or nothing at all as allowing benchmarks to do it brings back false results.
The fact you think this practice is fine baffles me TBH.
Anandtech,s view is not shared by samsung and I tend to agree with samsung. The gpu does not perform at 533 solely when a benchmark is detected it can under normal conditions for certain apps perform at this speed a benchmark being one of these apps.
However it cannot, because of heat issues, perform at such in full screen mode for prolonged periods of time. This limitation does not mean it therefore is solely for benchmarks if that was the case the gpu would not be capable of such speeds full stop other than in benchmarks.
I can see where anandtech have drawn their conclusion but they have obtained this is the context of a benchmark and not sought to properly research outside of this.
As stated in their original article.
"Samsung awesomely exposes the current GPU clock without requiring root access. Simply run the following command over adb and it’ll return the current GPU frequency in MHz:
Let’s hope this doesn’t get plugged, because it’s actually an extremely useful level of transparency that I wish more mobile platform vendors would offer.
Running any games, even the most demanding titles, returned a GPU frequency of 480MHz - just like @AndreiF alleged. Samsung never publicly claimed max GPU frequencies for the Exynos 5 Octa (our information came from internal sources), so no harm no foul thus far."
There is now an update from them based on the samsung statement in which they concede that this gpu speed is also available for limited uses and is available as samsung has said.
I however do agree that is could be misleading if you really are reliant on benchmarks and gaming is your thing, but there is a big difference between what is happening and deliberately and solely setting the gpu to perform this way in the driver solely for benchmarking as in the days of nvidia and ati past.
The solution anandtech would like to see is making such speed user definable but they also concede that samsung have limited this for a good thermal reason.
Anandtech,s view is not shared by samsung and I tend to agree with samsung. The gpu does not perform at 533 solely when a benchmark is detected it can under normal conditions for certain apps perform at this speed a benchmark being one of these apps.
However it cannot, because of heat issues, perform at such in full screen mode for prolonged periods of time. This limitation does not mean it therefore is solely for benchmarks if that was the case the gpu would not be capable of such speeds full stop other than in benchmarks.
I can see where anandtech have drawn their conclusion but they have obtained this is the context of a benchmark and not sought to properly research outside of this.
As stated in their original article.
"Samsung awesomely exposes the current GPU clock without requiring root access. Simply run the following command over adb and it’ll return the current GPU frequency in MHz:
Let’s hope this doesn’t get plugged, because it’s actually an extremely useful level of transparency that I wish more mobile platform vendors would offer.
Running any games, even the most demanding titles, returned a GPU frequency of 480MHz - just like @AndreiF alleged. Samsung never publicly claimed max GPU frequencies for the Exynos 5 Octa (our information came from internal sources), so no harm no foul thus far."
There is now an update from them based on the samsung statement in which they concede that this gpu speed is also available for limited uses and is available as samsung has said.
I however do agree that is could be misleading if you really are reliant on benchmarks and gaming is your thing, but there is a big difference between what is happening and deliberately and solely setting the gpu to perform this way in the driver solely for benchmarking as in the days of nvidia and ati past.
The solution anandtech would like to see is making such speed user definable but they also concede that samsung have limited this for a good thermal reason.
So how to you explain the white list which includes benchmarking tools? It's not just an accident that these run at a faster clock speed, Samsung has intentionally set it up this way.
Have you been waiting, not read this properly yet but I will give you my initial take as you have been waiting so patiently.
Don't see it, I can't imagine samsung or any phone manufacturer doctoring a phone to run a benchmark quicker just doesn't seem realistic or likely to me.
There is such a small market that would notice or even care about such things it would be more effort than its worth. The days of Nvidia and ATI doing this are completely different.
It's not denial at all. It's called debating. He simply like i doesnt agree this was an intentional act.
There is absolutely no point in Samsung doing this to fudge a number that makes no odds and is normally completely wrong anyway and mean absolutely nothing to about 99% of people!!
I await to be called a Samsung fanboy now i guess.
I agree, I don't understand why samsung would do this. You have to admit, coding a string called "benchmarkbooster" does raise suspicion.
"As far as synthethic benchmarks are concerned there isn't too much of a difference between the Octa and Snapdragon 600 versions of Samsung's best droid yet. We suspect this is on purpose - "
"Samsung wanted to guarantee that devs can expect the exact same level of performance form the Galaxy S4, regardless of which version the user has"
I do agree in an ideal world the benchmark would be better at 480, but to some degree are samsung to be held responsible for the accuracy of benchmarks? Those coding the benchmark should perhaps ensure they are accurate.
However the fact still remains that the phone is capable of such speeds and it is not just the domain and purpose of benchmarking. However samsung are usually pretty quick to react to such things and will probably close it off, as anandtech said it was only because samsung had made this open and transparent in the first place.
Having said all that it could be some fix attempted by samsung but I, looking at all the facts just don't see it based on the possible reward and trouble involved in doing this.
I do agree in an ideal world the benchmark would be better at 480, but to some degree are samsung to be held responsible for the accuracy of benchmarks? Those coding the benchmark should perhaps ensure they are accurate.
Samsung choose when it can run full speed and they allow it to run in those benchmarks to give better figures. To my mind that is up to Samsung to decide how it works.
Comments
They have intentionally made benchmarks look better than real world usage, yes that is what I think.
But then the benchmark tool may trigger it which isn't Samsung's issue. A benchmark test is over with in seconds normally. A game can be run for hours.
Nothing to do with a benchmark being detected as far as that statement goes the phone is capable of such gpu performance and it is not limited solely to a benchmark as the statement clearly says.
Your statement would imply that the drivers have been altered purely for such benchmarks, that is not in line with the statement
I have read the article and as I said from the outset I just don't see it. I understand where they may have reached this conclusion but I don't blindly follow or believe something because its posted.
I do however think that the performance may not be indicative of real world game performance as the gpu is not capable over prolonged periods but this is a long way from doctoring benchmark results. As samsung state the gpu is capable of such speed.
Sorry I am now confused is it underclocking for games. Apps which do not drain the battery get full GPU
According to Anandtech the phones GPU runs at a standard 480MHz but when certain benchmarks are detected (a whitelist if you will) it overclocks this to 523MHz. The reason it doesn't run at 523MHz all the time is unknown but it would probably cause thermal issues if used for extended periods.
As mentioned by Anandtect
IMO Samsung should either allow games to also run at 523MHz (which wont happen as samsung said it cannot run that for an extended period of time) or nothing at all as allowing benchmarks to do it brings back false results.
The fact you think this practice is fine baffles me TBH.
However it cannot, because of heat issues, perform at such in full screen mode for prolonged periods of time. This limitation does not mean it therefore is solely for benchmarks if that was the case the gpu would not be capable of such speeds full stop other than in benchmarks.
I can see where anandtech have drawn their conclusion but they have obtained this is the context of a benchmark and not sought to properly research outside of this.
As stated in their original article.
"Samsung awesomely exposes the current GPU clock without requiring root access. Simply run the following command over adb and it’ll return the current GPU frequency in MHz:
Let’s hope this doesn’t get plugged, because it’s actually an extremely useful level of transparency that I wish more mobile platform vendors would offer.
Running any games, even the most demanding titles, returned a GPU frequency of 480MHz - just like @AndreiF alleged. Samsung never publicly claimed max GPU frequencies for the Exynos 5 Octa (our information came from internal sources), so no harm no foul thus far."
There is now an update from them based on the samsung statement in which they concede that this gpu speed is also available for limited uses and is available as samsung has said.
I however do agree that is could be misleading if you really are reliant on benchmarks and gaming is your thing, but there is a big difference between what is happening and deliberately and solely setting the gpu to perform this way in the driver solely for benchmarking as in the days of nvidia and ati past.
The solution anandtech would like to see is making such speed user definable but they also concede that samsung have limited this for a good thermal reason.
So how to you explain the white list which includes benchmarking tools? It's not just an accident that these run at a faster clock speed, Samsung has intentionally set it up this way.
Just the in denial response I was expecting.
It's not denial at all. It's called debating.
If Samsung have done this deliberately and it wasn't just left over coding from early versions then its a bloody stupid thing to do.
Not that it would make any difference to real life use of the phone gaming or otherwise.
I agree, I don't understand why samsung would do this. You have to admit, coding a string called "benchmarkbooster" does raise suspicion.
http://www.gsmarena.com/samsung_galaxy_s4_i9500_vs_i9505-review-930p2.php
"As far as synthethic benchmarks are concerned there isn't too much of a difference between the Octa and Snapdragon 600 versions of Samsung's best droid yet. We suspect this is on purpose - "
"Samsung wanted to guarantee that devs can expect the exact same level of performance form the Galaxy S4, regardless of which version the user has"
What you mean a reasoned response based on my intuition rather than the fact whether I own the phone or not.
However the fact still remains that the phone is capable of such speeds and it is not just the domain and purpose of benchmarking. However samsung are usually pretty quick to react to such things and will probably close it off, as anandtech said it was only because samsung had made this open and transparent in the first place.
Having said all that it could be some fix attempted by samsung but I, looking at all the facts just don't see it based on the possible reward and trouble involved in doing this.
Samsung choose when it can run full speed and they allow it to run in those benchmarks to give better figures. To my mind that is up to Samsung to decide how it works.