• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Real world speeds from T-Mo's 42Mbps network

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Both the Amaze and the SGS2 max out at 42Mbps, in theory. In reality, you have to live in one of the few markets, climb on the roof at 3 am to get the best reception and least network usage and run a test that will still net you 40% of that capacity. And anytime you get a triple digit latency the advantage of the high speed is heavily negated by it.

Compare land line based wifi or reverse USB tethering. I would rather be on a DSL based 5Mbps wifi connection than a 10Mbps HSPA+ connection any day.

The smartphone bottleneck becomes obvious after browsing on a 2006 desktop with a 5Mbps DSL connection and comparing it to smartphone browsing on 10Mbps+ connections. HSPA+ and LTE have caught up with the land-line counterparts in terms of speed, but the hardware has long ways to go. My point is that our hardware is lagging behind and prevents us from truly taking advantage of the existing networks let alone those 20Mbps+ connections that carriers want us to buy. Network speed isn't the problem, it's the bandwidth cost, overage risk, carrier TOS and network capacity issues, along with present hardware inability to take full advantage of the bandwidth.

I'll take a better SoC that I benefit from constantly over a questionable network speed upgrade any day. Especially because it's not available in my market.
Less than 1/3rd of the market is covered with the 4G crap anyway and even if you happen to live in one, your precise location may prevent you from getting good speeds.

Hopefully they won't have to compromise performance again. Probably all 2012 T-Mobile phones will have 42 Mbps chipsets anyway. Then we can all go into that parking lot next to the tower and hit the cap in one day and enjoy the throttled speeds for the rest of the month.

Agree to disagree, I guess. I would rather have networks focus on improving coverage, latency and network capacity than speed bumping and advertising them. I don't even want to know how much $$$ telcos wasted on "blazing fast" hype ad campaigns. They aren't necessarily mutually exclusive but we have seen what network overloads have done in the past.


By summer 2012, high end phones with 1280x720p displays should be widespread. Streaming a 720p clip with HQ sound will net you anywhere between 4-8 Mbps depending on compression, frame rate and content. Today phone modems can easily handle that speed on HSPA+, LTE and wifi networks but the CPU and GPU can't process the content.
 
Last edited:
Verizon 4G LTE, tethered:
UCLyA.png
 
Course, then the myriad of businesses, and jobs, created based around smartphones fall flat, innovation is stifled, and carriers upgrade their networks at an even slower pace, phone technology slows down because people don't have a need to upgrade, etc.

Metered billing = bad for the consumer. Good for the carrier, but no one else.


When the carriers see an increasing number of users eating 10GB+/month and the average of everyone else increasing at a rate that's faster than bandwidth it's a no brain'r for them -- either they do something to curb data use (caps) or they increase the cost to cover greatly expanding the number of cell sites.

The second option isn't really an option in many cases as there are many places they just can't expand in that way. Metered billing will put a halt to the data hogs as they will wind up paying 2X-10X or more and there's no way to root or jail break around it. The fact that you can't root around it is the main thing that will drive the carriers to metered billing.

I expect to see metered billing within 5 years...


Brian
 
Technically you're right, but the obvious software analogy is that you compress the data traveling over the network, essentially making it appear faster. On the other hand, this is already done for a lot of data (e.g. audio and video) so in a way we've already optimized that part of the software. I suppose one could look into compressing other types of data that aren't currently being compressed. If the SoC had dedicated hardware to decompress it, there wouldn't be too much of a penalty to speed/performance. Probably not worthwhile though.

What makes you think the data isn't being compressed now?


Brian
 
In reality of course, but you assume that every network is automatically over used. I don't have any issues with slow downs on t-mobile. Also you assume that a theoretically faster CPU is going to be noticeably faster in reality. Getting a better Quadrant score doesn't necessarily mean a more responsive device.

Yes, a faster network is always a good thing but as I said before, the rate at which data usage is increasing is faster than the rate that bandwidth is increasing so sooner or later, unless something is done to put a halt to data hogs, the network will slow to a crawl. Sure, there are places with fewer people hitting a cell site and places with fewer data hogs but that's just one end of the spectrum with most of us somewhere towards the other end.

And having a faster CPU/GPU will not guarantee better performance but it does make it easier to achieve it. In addition to processing HD video streams, gaming is very CPU/GPU demanding so if you're a gamer (I'm not) then a faster SoC will almost certainly be a good thing.


Brian
 
When the carriers see an increasing number of users eating 10GB+/month and the average of everyone else increasing at a rate that's faster than bandwidth it's a no brain'r for them -- either they do something to curb data use (caps) or they increase the cost to cover greatly expanding the number of cell sites.

The second option isn't really an option in many cases as there are many places they just can't expand in that way. Metered billing will put a halt to the data hogs as they will wind up paying 2X-10X or more and there's no way to root or jail break around it. The fact that you can't root around it is the main thing that will drive the carriers to metered billing.

I expect to see metered billing within 5 years...


Brian
If metered billing comes into play, that means you're also expecting tethering/hotspot plans to go the way of the dodo bird.
I don't see the US carriers giving up on their juicy tethering/hotspot plans anytime soon.

How will metered billing take effect?
Will it be like electricity where it based flat across the board rate on the kwh used or are the US carriers going to create an iAnal version like we already see today?
Example of iAnal: 100-300MB data plans being offered by the major carriers ranges from $15-20/month...2GB is only $5-10 more.
That's not metered billing...That's called iAnal.
Lets not even talk about the ridiculous "overage" fees on those 200-300MB data plans being offered by Verizon/AT&T compared to the 2GB data plans.

If I could be charged an across the board flat rate that would be great so one doesn't have to worry about which group of data users to belong to.
I doubt that US carriers would be as straight forward with metered billing as you think they'll be. They'll most likely implement some iAnal version that's even worse than what we already have today.
 
I get simular results on Tmos 3g on my vibrant.

You get pings in the 50s on TMO 3G? Got a screen shot?

I expect to see metered billing within 5 years...

Brian

When you say metered billing, I assume you mean something along the lines of X amount of dollars per GB and not the current common 2GB caps? Paying per GB used pretty much kills everything that people use a smartphone for, as the more you use it, the more expensive it'll be. Right now, these devices are getting into the hands of more and more people, driving millions of new customers and their revenue to the carriers. And the services that exist on those phones as well. If the bill increases to the level you're suggesting, the system collapses. There's no need to have a high end phone, with high res screen and powerful SoC, and blazing fast data, if watching a single 45min TV episode adds a 100 bucks to your bill. Having a monthly recurring bill that fluctuates wildly every month is obscene.

Carriers will be shooting themselves in the foot by jacking prices up that much, not to mention causing dozens of other businesses to collapse.

On data caps, I'm not opposed to them. But, I am opposed to the extremely low 2GB that carriers are making the standard. Set the cap at 5-7GB, and throttle speeds to about 1/4th of normal when the user exceeds that until their next billing cycle starts. Also, include tethering with this pool of data. Tethering right now is a gimmick for most people, only really only useful for some corporate uses. Give people a reason to tether their phone and use the data, and they'll have a reason to pay for a plan with a higher data cap. This method doesn't stifle the user, doesn't impede the network, doesn't hurt the businesses that depend on the user actually using their phone, and as a bonus, it encourages people to use services and applications because they know what the bill will be each month.
 
What makes you think the data isn't being compressed now?

Most of it already is. Any audio or video is compressed. Images are compressed. The rest of it is mostly text, which doesn't amount to much, so it's really not worth the trouble and since there's no dedicated hardware to decompress is as there is for audio, video, etc. it would just drain the battery more.
 
Yes, a faster network is always a good thing but as I said before, the rate at which data usage is increasing is faster than the rate that bandwidth is increasing so sooner or later, unless something is done to put a halt to data hogs, the network will slow to a crawl. Sure, there are places with fewer people hitting a cell site and places with fewer data hogs but that's just one end of the spectrum with most of us somewhere towards the other end.

And having a faster CPU/GPU will not guarantee better performance but it does make it easier to achieve it. In addition to processing HD video streams, gaming is very CPU/GPU demanding so if you're a gamer (I'm not) then a faster SoC will almost certainly be a good thing.


Brian

I live about three blocks from Microsoft's campus, this isn't exactly a low tech area lacking in data hogs. Conversely, as T-Mobile continues to improve their HSPA+ network, my throughputs are staying as fast or even getting faster at times.

Look, in an ideal situation, you'd have both the fastest SoC and network, but that isn't an option here. Everyone has their preferences, and my preference would be the faster data speed.
 
Put me down for the better SOC. My Exynos SGS2 can play almost every mkv I have (in the default player). My friend's Evo3D could only play a few even with a better more compatible player. The Exynos's decoder wins till Tegra 3.
 
You get pings in the 50s on TMO 3G? Got a screen shot?

I swapped phones with my gf who has a HSPA+ capable G2:

91889070.png


91890581.png


Now, granted, that's "4G", but on my Vibrant (not 4G) my numbers really aren't much worse. Downloads ~ 5Mb, pings around 60-70.
 
Last edited:
I swapped phones with my gf who has a HSPA+ capable G2:

91889070.png


91890581.png


Now, granted, that's "4G", but on my Vibrant (not 4G) my numbers really aren't much worse. Downloads ~ 5Mb, pings around 60-70.

Know that you are an exception. 8 out of 10 phones can't get those low pings. If I had to gauge the average of all T-Mo pings I saw, your results are twice as good, as the average both speed and latency.
 
You get pings in the 50s on TMO 3G? Got a screen shot?



When you say metered billing, I assume you mean something along the lines of X amount of dollars per GB and not the current common 2GB caps? Paying per GB used pretty much kills everything that people use a smartphone for, as the more you use it, the more expensive it'll be. Right now, these devices are getting into the hands of more and more people, driving millions of new customers and their revenue to the carriers. And the services that exist on those phones as well. If the bill increases to the level you're suggesting, the system collapses. There's no need to have a high end phone, with high res screen and powerful SoC, and blazing fast data, if watching a single 45min TV episode adds a 100 bucks to your bill. Having a monthly recurring bill that fluctuates wildly every month is obscene.

Carriers will be shooting themselves in the foot by jacking prices up that much, not to mention causing dozens of other businesses to collapse.

On data caps, I'm not opposed to them. But, I am opposed to the extremely low 2GB that carriers are making the standard. Set the cap at 5-7GB, and throttle speeds to about 1/4th of normal when the user exceeds that until their next billing cycle starts. Also, include tethering with this pool of data. Tethering right now is a gimmick for most people, only really only useful for some corporate uses. Give people a reason to tether their phone and use the data, and they'll have a reason to pay for a plan with a higher data cap. This method doesn't stifle the user, doesn't impede the network, doesn't hurt the businesses that depend on the user actually using their phone, and as a bonus, it encourages people to use services and applications because they know what the bill will be each month.

if i had to guess, at the rate flash is increasing we won't need insanely fast wireless speeds soon. you can just carry enough content with you on your device
 
if i had to guess, at the rate flash is increasing we won't need insanely fast wireless speeds soon. you can just carry enough content with you on your device

Fat load of good that does me when I'm trying to watch HQ Youtube vids or stream a sports game, or watch Netflix, etc. 😉
 
Know that you are an exception. 8 out of 10 phones can't get those low pings. If I had to gauge the average of all T-Mo pings I saw, your results are twice as good, as the average both speed and latency.

I am in houston and my ping is usually 75ish. but my download range from 3-7 usually 3-4.
 
Back
Top