Didn't expect the gap to be that big.

[chanchan05]

If we assume the tests are fair the only conclusion we can draw is the one phone is slightly faster at certain functions. Suggesting that the extra ram is the source of difference is a good guess but we don't really know for sure. Extra ram only improves speed if you don't have enough ram to begin with. I'm not saying that the extra ram isn't helping, however 4 GB of RAM on a phone is pretty powerful.

If you watched the test, the gap widened because in the second lap where the apps they opened in the 1st lap were reopened, the Note 8 kept all apps running in the background, while the v30 reloaded some apps, so it was incapable of holding them all in background. So the RAM was insufficient for it, or it has bad RAM management.

 

[nizmoz]

It''s way more credible than your video. EOD.

 

[Mike Dee]

If you watched the test, the gap widened because in the second lap where the apps they opened in the 1st lap were reopened, the Note 8 kept all apps running in the background, while the v30 reloaded some apps, so it was incapable of holding them all in background. So the RAM was insufficient for it, or it has bad RAM management.

We can only assume it is one of those two options since we don't really know. In the end, does the difference in speed really matter?
I don't own either device so for me it doesn't, however I don't see it as the main reason to buy a phone. I guess to some people it's important because here we are.

 

[Will In SF]

These speed tests are of no importance to me, just like I have little interest in benchmarks.

 

[irvine752]

1. Actually if you want to show real life performance you aren't going to power cycle. Not everyone power cycles the phone before every usage.

2. All apps used in the test obviously loaded when opened, hence none of them were already sitting in the background for either phone.

3. And power cycling would make the test irrelevant when you want the test to show how fast the phone could be in real life conditions. In real life several apps would have tried to run in the background already and the phone would have several apps cached, not empty. What you're looking for is a drag race where we strip everything down to minimum and race with just the engines. That race is useless. We need to race with everything on board. If the backseats on the other car is too heavy making it lose to a coupe, that's too bad. The test that happened is basically if it would be faster to carry 6 apples at a time than 4.
You only need the power cycle if you want to see the absolute performance of the device. If that's not what you're after, then the power cycling skews the test to be erroneous. If there's a memory leak, then the power cycling would hide it's effect on performance in the speed test, but in real usage the phone would be slower.

4. The Note 8 based on previous tests, we know is performing as expected, slightly worse even. And we can't say Phonebuff is biased when last time it did such tests, Samsungs were the worst performers (last years S7 Snapdragons). The S8 lost to the HTC U11 and G6 if I remember correctly. This is perhaps the first time LG got beat that I recall. And this has the Note 8 which has been in house for several weeks going up against a new acquired v30. Wouldn't it be the Note 8 supposedly be at a disadvantage if no power cycling happened? The Phonebuff speed test was never about absolute hardware speed.

5. Let's just see how the v30 performs in subsequent tests. It should have times close to what occured now.

The power cycle is not to show real life performance. It's a virtual starting line just so everybody can have confidence that nothing else was sitting in RAM & taking up processing power.

Whether the Apps loaded or not that's not the point. He could have easily had other different Apps or even worse a high intensive game running & sitting in the RAM (only on the V30) when he started the test. That alone could have easily slowed down the perfomance of the V30. That's why when you do these speed tests you try to eliminate any room for bias. Clearing the RAM certainly helps the cause since the premise of the video was to determine if additional RAM made any difference or not.

 

[chanchan05]

The power cycle is not to show real life performance. It's a virtual starting line just so everybody can have confidence that nothing else was sitting in RAM & taking up processing power.

Whether the Apps loaded or not that's not the point. He could have easily had other different Apps or even worse a high intensive game running & sitting in the RAM (only on the V30) when he started the test. That alone could have easily slowed down the perfomance of the V30. That's why when you do these speed tests you try to eliminate any room for bias. Clearing the RAM certainly helps the cause since the premise of the video was to determine if additional RAM made any difference or not.

Again it would skew the test because clearing the RAM clears anything the LG preloads on its own. Which you should not do. Let the LG load everything it wants even if it ends up loading more than the Note 8. If it performs worse because of that, then it's LG's fault for poor software. Which is also a reason why I find doing a speed test right after booting or RAM clearing is a stupid idea. It's not gonna perform like that 3hrs later. You're looking at highest possible speed there, not speed on regular usage.

We're not starting from a zero line here. We're starting from whatever LG and Samsung decided to do. If Samsung's software tells the phone to keep 2GB free and LG's software tells the phone to eat all the RAM it wants, then clearing the RAM creates a false result. It's LG's own fault if it ends up handicapped because there's more stuff in the RAM.

It again comes down to the race car analogy. Two cars with same engine, but one is heavier. It's like you're saying we should trim the other car so it has the same weight as the other for it to be fair. Which we don't want to do. They should be raced as is, because people won't be buying the trimmed down car, they're buying the fully loaded car.

It's either you accept that or claim that the test was rigged in favor of the Note, where we'll have enough proof later on when more v30 videos from Phonebuff later where we'll see if the v30 performs at the same rate.

 

[irvine752]

Again it would skew the test because clearing the RAM clears anything the LG preloads on its own. Which you should not do. Let the LG load everything it wants even if it ends up loading more than the Note 8. If it performs worse because of that, then it's LG's fault for poor software. Which is also a reason why I find doing a speed test right after booting or RAM clearing is a stupid idea. It's not gonna perform like that 3hrs later. You're looking at highest possible speed there, not speed on regular usage.

We're not starting from a zero line here. We're starting from whatever LG and Samsung decided to do. If Samsung's software tells the phone to keep 2GB free and LG's software tells the phone to eat all the RAM it wants, then clearing the RAM creates a false result. It's LG's own fault if it ends up handicapped because there's more stuff in the RAM.

It again comes down to the race car analogy. Two cars with same engine, but one is heavier. It's like you're saying we should trim the other car so it has the same weight as the other for it to be fair. Which we don't want to do. They should be raced as is, because people won't be buying the trimmed down car, they're buying the fully loaded car.

It's either you accept that or claim that the test was rigged in favor of the Note, where we'll have enough proof later on when more v30 videos from Phonebuff later where we'll see if the v30 performs at the same rate.

A power cycle clears RAM that would a reboot for a phone. When you reboot your phone both manufactures still load their respective software. It's really not that confusing. Did you not see the second video?

If you want to stick to your race car analogy, this would be like having a drag race but with one car starting a couple feet ahead which would be unfair.

 

[chanchan05]

A power cycle clears RAM that would a reboot for a phone. When you reboot your phone both manufactures still load their respective software. It's really not that confusing. Did you not see the second video?

If you want to stick to your race car analogy, this would be like having a drag race but with one car starting a couple feet ahead which would be unfair.

The OS does not load everything up on boot. Some things are only loaded up once a different process has called them. Or when the prefetch software kicks in.

Whether they stay on or not is the system's own fault.

And whether or not that is the case, Phonebuff reviews come from a fresh reboot anyway and are connected to the same network. So all your arguments are invalid. They stopped showing the bootup some time ago because they felt that bootup takes too much time and majority of people hardly care about bootup times anymore. You can go back maybe 3 years ago when they stopped including the bootup time in their obstacle course test.

The only argument you can make is whether they cheated or not regarding the phones being freshly booted or not.

 

[Rumblee1]

This is making me rethink getting the V30

I received my v30 yesterday. Got it all set up and honestly, it's super fast. No hangup whatsoever. Remember, as stated above, real world usage is incredible and quite different than a side-by-side comparison. Lg knocked it out of the park with the v30. And wait till you use the quad dac. WOW

 

[chanchan05]

I received my v30 yesterday. Got it all set up and honestly, it's super fast. No hangup whatsoever. Remember, as stated above, real world usage is incredible and quite different than a side-by-side comparison. Lg knocked it out of the park with the v30. And wait till you use the quad dac. WOW

The only reason the v30 lost that way was because of RAM and RAM handling difference. In the first part they were pretty neck and neck on speed. So there should be no issues in real world usage. You're not gonna switch between 10 apps in 2min anyway. So this speed test really shouldn't be a basis if you're buying or not. It's just bragging rights.

Edit: we all knew the N8 has more RAM even becore the test was done anyway.

 

[chriskwarren]

These things are neat experiments to show how each phone handles somewhat extreme scenarios, but I doubt this would be very noticeable to most folks. That said, good to see the N8 handle that workload...impressive.

 

[irvine752]

The OS does not load everything up on boot. Some things are only loaded up once a different process has called them. Or when the prefetch software kicks in.

Whether they stay on or not is the system's own fault.

And whether or not that is the case, Phonebuff reviews come from a fresh reboot anyway and are connected to the same network. So all your arguments are invalid. They stopped showing the bootup some time ago because they felt that bootup takes too much time and majority of people hardly care about bootup times anymore. You can go back maybe 3 years ago when they stopped including the bootup time in their obstacle course test.

The only argument you can make is whether they cheated or not regarding the phones being freshly booted or not.

I agree the OS does not load everything up on boot up. That's precisely the point, it guarantees a cold start of the Apps without any residue in the system. When the CPU starts executing instructions, the main memory (RAM) simply stores those computations. The Android OS does not support swapping, which would mean that when your memory starts running out or pushing for its upper limit, the CPU starts asking other processes to release their resources (garbage collection). The less memory you have the the longer you have to wait to start a new process, which would create a delayed response. The summation of these delayed responses is what you're seeing in the first video. It's basic CPU architecture, hence the need for a power cycle. Looking at the facts right now (without any assumptions from years ago), he clearly did not show or do a power cycle otherwise the results would have been similar to the second video. It also really doesn't take long to do one either, adding an extra minute or two to the video won't make a difference.

Just so you know, contrary to popular belief (without taking into consideration the software overlays TouchWiz vs LG UX 6.0) main memory doesn't really effect the launch speed of apps. Just like a computer, the launch speed of applications are actually dependent on the speed of the CPU & it's corresponding secondary storage, which would be the cache/flash memory. Once an application is in memory it's already running so the dependency would be on how long it takes the process to go from secondary memory to main memory. Keep in mind the CPU on the V30 is clocked higher which would be the slight advantage shown in the second video.

Having more RAM (Note 8) simply means better task-switching & the added advantage of running more complicated apps or even opening larger/more complex files. A faster CPU (V30) means having a faster turnaround between input and response (less lag). CPU-bound apps would also benefit with better background processing. In short, both are equally important; your own personal usage pattern will determine which is more valuable for you. If you generally stick with in a single program and you're sensitive to input-output latency then having faster CPU will be more valuable for you. If you constantly task switch between many different apps or if you need to run complex memory-hungry apps to open complex large files, then having extra RAM is going to be more valuable to you.

 

[chanchan05]

I agree the OS does not load everything up on boot up. That's precisely the point, it guarantees a cold start of the Apps without any residue in the system. When the CPU starts executing instructions, the main memory (RAM) simply stores those computations. The Android OS does not support swapping, which would mean that when your memory starts running out or pushing for its upper limit, the CPU starts asking other processes to release their resources (garbage collection). The less memory you have the the longer you have to wait to start a new process, which would create a delayed response. The summation of these delayed responses is what you're seeing in the first video. It's basic CPU architecture, hence the need for a power cycle. Looking at the facts right now (without any assumptions from years ago), he clearly did not show or do a power cycle otherwise the results would have been similar to the second video. It also really doesn't take long to do one either, adding an extra minute or two to the video won't make a difference.

Just so you know, contrary to popular belief (without taking into consideration the software overlays TouchWiz vs LG UX 6.0) main memory doesn't really effect the launch speed of apps. Just like a computer, the launch speed of applications are actually dependent on the speed of the CPU & it's corresponding secondary storage, which would be the cache/flash memory. Once an application is in memory it's already running so the dependency would be on how long it takes the process to go from secondary memory to main memory. Keep in mind the CPU on the V30 is clocked higher which would be the slight advantage shown in the second video.

Having more RAM (Note 8) simply means better task-switching & the added advantage of running more complicated apps or even opening larger/more complex files. A faster CPU (V30) means having a faster turnaround between input and response (less lag). CPU-bound apps would also benefit with better background processing. In short, both are equally important; your own personal usage pattern will determine which is more valuable for you. If you generally stick with in a single program and you're sensitive to input-output latency then having faster CPU will be more valuable for you. If you constantly task switch between many different apps or if you need to run complex memory-hungry apps to open complex large files, then having extra RAM is going to be more valuable to you.

And you point out exactly why a restart is not needed or even wanted for a speed test that seeks to emulate real life condition. In real life you will not be using the phone fresh from a restart always. You will be using it after it has been sitting around executing background processes for some time already. So a speed test of two phones that did not go a reboot is a more valid representation of real world speed as compared to a speed test done on a freshly power cycles device. We are not looking for the max speed the phones are capable of. We are looking for the usual speed we can expect after they have been sitting around executing whatever process they want to execute on their own. I don't see why it's so hard for you to see that point. You keep on asking for a cold restart which will show max possible speed when I keep repeating that's not the type speed results we're looking for. We're looking for the speed under normal usage circumstances, and that means phones that haven't been rebooted. And that's the case for non rebooted apps. But you insist that Phonebuff probably cheated and did not power cycle.

And second, the second video does not invalidate the first video. The second video gave more points to the LG on a per app basis, while the 1st video gave the score to Samsung on an accumulated time basis. Two very different scoring systems. For example, in the first video by Phonebuff it would not matter if the LG opened 10 apps faster than the Samsung by one tenth of a second if it was 2 seconds behind in one app, giving Samsung the win. That would be a 10-1 score against the Samsung in the scoring system used in the second video. See the difference? Even in the second video during the task switching part the Samsung was faster than the LG, best seen probably on Asphalt 8. Also in apps where LG won in opening, it was pretty close. But in say the Youtube app, LG was half a second or more slower. There were also some instances where to my eyes either the LG or Samsung was faster, but the reviewer called it for the other device. Second, side tests that are performed with one hand for either phone are iffy because both hands will not be always moving at the same time no matter what you think. There is always a fraction of a second difference between them mainly because of how our brain is wired. The dominant hand will always be faster than the other. And you cannot focus on two things at the same time, hence as I said the inconsistency on points given as I said. The phonebuff style test were each phone goes through the app obstacle course and is timed independently of the other device is more valid than side by side test by simple virtue of human anatomy.

So your arguments that the 2 videos show different results do not hold as well. It's mainly 2 different ways of scoring with one of them even more prone to error.

Let's put it this way. After your phone has been sitting on your desk all night while you sleep, and you wake up wanting to play a game, say SIM City, do you reboot or just open the game? You just open the game. That's a real life scenario. So a fresh reboot speed is farther from real world usage speeds.

 

[irvine752]

And you point out exactly why a restart is not needed or even wanted for a speed test that seeks to emulate real life condition. In real life you will not be using the phone fresh from a restart always. You will be using it after it has been sitting around executing background processes for some time already. So a speed test of two phones that did not go a reboot is a more valid representation of real world speed as compared to a speed test done on a freshly power cycles device. We are not looking for the max speed the phones are capable of. We are looking for the usual speed we can expect after they have been sitting around executing whatever process they want to execute on their own. I don't see why it's so hard for you to see that point. You keep on asking for a cold restart which will show max possible speed when I keep repeating that's not the type speed results we're looking for. We're looking for the speed under normal usage circumstances, and that means phones that haven't been rebooted. And that's the case for non rebooted apps. But you insist that Phonebuff probably cheated and did not power cycle.

I have no problems with tests that emulate real life performance. I only have a problem with ones that are not done properly like your favorite video. Like I said before, you need to level the starting point. You can do this by making sure both phones are running with same amount of used memory or do a power cycle to clear the RAM. Neither of those were done in the first video, hence it can't be trusted. If the test is started with different amounts of used main memory/RAM available, the test would be skewed. Using a similar CPU, a device with say 10% used memory will always be faster than one with 90% used memory. Yes the tester could have cheated & easily had stuff running in background for the V30 & using up almost all it's RAM which would to lead to the garbage collection & the delayed reponse I mentioned before.

 

[chanchan05]

I have no problems with tests that emulate real life performance. I only have a problem with ones that are not done properly like your favorite video. Like I said before, you need to level the starting point. You can do this by making sure both phones are running with same amount of used memory or do a power cycle to clear the RAM. Neither of those were done in the first video, hence it can't be trusted. If the test is started with different amounts of used main memory/RAM available, the test would be skewed. Using a similar CPU, a device with say 10% used memory will always be faster than one with 90% used memory. Yes the tester could have cheated & easily had stuff running in background for the V30 & using up almost all it's RAM which would to lead to the garbage collection & the delayed reponse I mentioned before.

The starting point cannot be level if the system itself is the one that loaded up all the excess stuff in the first place.

For example, let's say both phones were sitting for 8 hours after bootup, but the LG system loaded up 90% of RAM, but the Samsung system only loaded up 60% of RAM. The LG would of course be slower in that scenario, but the test would be valid because they are coming from the same starting point of being booted up several hours before. Making a bootup would not give them both the same starting point. It would hide the flaws in the LG memory management (example only), which would make the test invalid. So basically I'm saying the entire idea of doing a power cycle before a speed test is a wrong representation of actual world speed, if that is your argument. It doesn't matter if one device has more RAM used up than the other. It's the device's own fault for loading all of that. This is similar to how the Snapdragon S7 lost to the G5 last year in the Phonebuff video. The S7 loaded up more stuff onto memory and had troubles swapping. In that scenario it was the fault of the Samsung OS. And in this case, it's the fault of the LG OS. The test starting would different amounts of RAM available does not skew the test unless it was deliberately done (like deliberately loading apps onto the LG RAM before the test). However if it was the system itself that called those things to RAM on its own, it shot itself in the foot and it's no concern since it is simply reflective of how the phone OS was written to perform.

Your analogy of starting 2 meters ahead in a race is wrong. It's going to a race where one car has 3 passengers, and the other has only one. It's the other racer's own fault he brought passengers making his car heavier and imbalanced.

 

[irvine752]

The starting point cannot be level if the system itself is the one that loaded up all the excess stuff in the first place.

For example, let's say both phones were sitting for 8 hours after bootup, but the LG system loaded up 90% of RAM, but the Samsung system only loaded up 60% of RAM. The LG would of course be slower in that scenario, but the test would be valid because they are coming from the same starting point of being booted up several hours before. Making a bootup would not give them both the same starting point. It would hide the flaws in the LG memory management (example only), which would make the test invalid. So basically I'm saying the entire idea of doing a power cycle before a speed test is a wrong representation of actual world speed, if that is your argument. It doesn't matter if one device has more RAM used up than the other. It's the device's own fault for loading all of that. This is similar to how the Snapdragon S7 lost to the G5 last year in the Phonebuff video. The S7 loaded up more stuff onto memory and had troubles swapping. In that scenario it was the fault of the Samsung OS. And in this case, it's the fault of the LG OS. The test starting would different amounts of RAM available does not skew the test unless it was deliberately done (like deliberately loading apps onto the LG RAM before the test). However if it was the system itself that called those things to RAM on its own, it shot itself in the foot and it's no concern since it is simply reflective of how the phone OS was written to perform.

Your analogy of starting 2 meters ahead in a race is wrong. It's going to a race where one car has 3 passengers, and the other has only one. It's the other racer's own fault he brought passengers making his car heavier and imbalanced.

The excess stuff on boot up can be disregarded since it's intrinsic & won't affect the objectivity of test. The same holds true if you let the phones run for 8 hours after a boot up. You would just need to make sure that any external objects introduced to the memory space such as games & etc are fully closed on both phones (with nothing pending I/O or CPU bound) before you start your test. If any of those variables are not constant, you would lose the ideal of objectivity which would render the test invalid.

Your analogy is misguided. It's all about managing resources. The objects in memory space can be re-allocated to different reference points at any point in time or even killed to return the used up resources if needed. My lovely analogy was based on CPU latency.

I admire your persistence, but some of these computer architecture concepts are not simple. They are also not common since they are taught in advanced college level classes. You really need to get a good grasp on the fundamentals first before using such videos as proof.

 

[chanchan05]

The excess stuff on boot up can be disregarded since it's intrinsic & won't affect the objectivity of test. The same holds true if you let the phones run for 8 hours after a boot up. You would just need to make sure that any external objects introduced to the memory space such as games & etc are fully closed on both phones (with nothing pending I/O or CPU bound) before you start your test. If any of those variables are not constant, you would lose the ideal of objectivity which would render the test invalid.

Your analogy is misguided. It's all about managing resources. The objects in memory space can be re-allocated to different reference points at any point in time or even killed to return the used up resources if needed. My lovely analogy was based on CPU latency.

I admire your persistence, but some of these computer architecture concepts are not simple. They are also not common since they are taught in advanced college level classes. You really need to get a good grasp on the fundamentals first before using such videos as proof.

In the end it all boils down to you not believing that the first video was not from a fresh bootup because it wasn't shown. The fact is, because it wasn't shown, you'd rather dismiss the first video as rigged, despite no evidence pointing to it, and none of the following videos posted even disproves the results of the first video.

It doesn't matter what test was done. It all boils down to you refusing to believe the results of the first video on the basis that it could be rigged.

As I said, let's wait for more videos and we'll see reproducibility. The Note 8 performance was already shown to be reproducible in 3 instances.

And you didn't change my argument because none of what you said disproves it in saying that a power cycle is not needed. There just needs to be no external influence. You just refuse to believe that there was no external influence on the first test.