Page 10 of 16 FirstFirst ... 56789101112131415 ... LastLast
Results 181 to 200 of 304
  1. ink883's Avatar
    Posts
    872 Posts
    Global Posts
    883 Global Posts
    #181  
    After a hard reset and only installing pbench, my pbench data access score climbed to 807%

    After a restore from backupman, went back down :-(
    Visor --> Visor Platinum --> Treo 300 --> Treo 600 --> Treo 650 --> Treo 700p --> Treo 755p --> Treo 800w --> Palm Pre
  2. #182  
    A score of 1384 on pBench memory access is obviously from an overclocked or otherwise ROM-modified 650, and is therefore a dishonest number and will thankfully be disregarded by honest forum members.

    An honest result appears to be around 950, which shows less than a 20% difference from the 700p. This is nowhere near the 40%-50% slowdown that some here have claimed, and indeed my daily use bears out the fact that there are no slowdowns or strange pauses on my unit. This certainly does not preclude others from having this issue, but verification of such issue using falsified benchmark results doesn't do anybody any good.

    If Palm needs to fix something in the 700p, then fine -- bring it on. We will all benefit. But what is of NO BENEFIT at all, to any of us, is creating a rallying cry based upon falsified information. This was never an honest discussion to begin with because the original premise was fraudulently backed up with erroneous benchmark numbers obtained from a Treo 650 that had a custom ROM and was otherwise "tampered" with in order to coax artificially high benchmark scores.

    I have no problem having an intelligent discussion in an intelligent forum, but let's at least be clear what it is that we are talking about, which is a real world slow down of about 20% of the 700p compared to a 650. NOT 50%.

    20% speed differentials are tolerable by most, and completely unnoticed by many, so the massive slowdowns others have been discussing must clearly be caused by something else. There is also the question of the EVDO stack requiring additional processing power, and it may indeed be a COMBINATION of factors rather than just a single impact that is resulting in the overall slowdowns that some are suffering.

    I installed the Java modules available on the Palm site, and instantly my pBench memory access scores fell by 100 points. Deleting the Java apps brought my score back up. What is the explanation for this? Do some of these applications require system resources (besides storage space) that would result in a slowdown? Could this all simply be application related?
  3.    #183  
    Yup, the 700p is much slower than the 650. I find it very odd that some users aren't perceptive enough to notice the difference and claim their 700p's to be "lightning fast". Oh well. I guess we should leave this discussion to the power-users who use their Treos for more than playing Solitaire. lol

    We'll just see when those users are praising the newfound speed after their ROM update.
  4. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #184  
    Quote Originally Posted by SonnyS
    I have no problem having an intelligent discussion in an intelligent forum, but let's at least be clear what it is that we are talking about, which is a real world slow down of about 20% of the 700p compared to a 650. NOT 50%.
    In fact, the 20% slowdown from a stock 650 you and chckhbrt have shown is for data access times only, and other benchmarks which show processing power, graphics, and calculations (Speedy, VFSMark) show a stock 700p to be slightly faster than a stock 650. Therefore, even a claim that the 700p is 20% slower is not accurate without specifying WHERE it's slower, since there are many factors which contribute to the overall speed of a device. A claim that "all of the benchmarks prove that prove about about a 50% decrease in speed compared to Sprint 650" is a false claim:

    Quote Originally Posted by 2000 Man
    All of the benchmarks prove about a 50% decrease in speed compared to the Sprint 650. Speedy, pBench, VFSMark, TCPMP benchmark, etc.
    Thankfully for us, a little more accurate and truthful testing has revealed around a 20% slowdown in data access times if pBench is accurate. That means that something that took 1.2 seconds to access on a stock 650 now takes 1.44 seconds to access on the 700p. Ink is quite right: most people(and programs) would not be able to tell the difference between a 20% decrease in speed.

    I think Ink & Sonny, with their testing, have shown that there may be a problem with specific units or 3-party software that is slowing down the data access times more than 20%. That would explain why some have problems and some don't. MANY developers seem to have quite with their pants down, which is surprising considering developer's units must have been available, and Palm clearly made changes to the NVFS memory to make the 700p more stable. I have heard some mention a batch of 700p with "bad memory chips" which would also explain the inconsistencies between what people have experienced.
  5. ink883's Avatar
    Posts
    872 Posts
    Global Posts
    883 Global Posts
    #185  
    Quote Originally Posted by SonnyS
    An honest result appears to be around 950, which shows less than a 20% difference from the 700p.
    The difference established range from 22%-47%. People may not notice 20% but they certainly will notice 47%. I notice it, ptunes playback makes it noticeable, there are enough complaints regarding speed to make me think there are plenty of users experiencing more then a 40% slowdown.

    Again, I still say we are only talking about data access, not CPU speed or other things. I will still maintain Speedy and VFSMark are the wrong tools and the wrong measurements.

    There may be bad treos or it maybe a memory handling issue. Its odd that the more apps people install the more they are experiencing a slow down. Java shouldn't rob someone of 100 points, thatís unacceptable.
    Visor --> Visor Platinum --> Treo 300 --> Treo 600 --> Treo 650 --> Treo 700p --> Treo 755p --> Treo 800w --> Palm Pre
  6. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #186  
    Actually, Ink, you scored an 807, so the established range would be 15.1%-47%. I'm not sure if it's more apps installed, or certain ones that are having the effect. The only way to know for sure is install every app one by one, and then test data access speed after each install. Again, it's possible that a fix will have to come from the individual developers rather than from Palm.
  7. Rob_T's Avatar
    Posts
    328 Posts
    Global Posts
    340 Global Posts
    #187  
    This speed and performance issue stuff on the 700p is just frustrating.

    Why Palm can't take a Treo 650 and add some more RAM and EVDO (and call it a 700p) without screwing it up is beyond me.

    Why couldn't they (Palm) take a 650, leave the button layout alone, add some RAM, set it up for an EVDO connection and otherwise leave it alone so they maintain the stability and other gains they already worked through on the 650?
  8. ink883's Avatar
    Posts
    872 Posts
    Global Posts
    883 Global Posts
    #188  
    Quote Originally Posted by dcpmark
    Actually, Ink, you scored an 807, so the established range would be 15.1%-47%. I'm not sure if it's more apps installed, or certain ones that are having the effect. The only way to know for sure is install every app one by one, and then test data access speed after each install. Again, it's possible that a fix will have to come from the individual developers rather than from Palm.
    Yes, good call, keep in mind 807 was only after a hardreset and that went away soon after I restored my files. This is in line with the other thread about making a treo lightening fast (those users experienced only a temporary speed up).

    15% slowdown as the best case scenario isn't a good sign either. I certainly hope Palm didn't cheap out with slow memory modules.
    Visor --> Visor Platinum --> Treo 300 --> Treo 600 --> Treo 650 --> Treo 700p --> Treo 755p --> Treo 800w --> Palm Pre
  9. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #189  
    Quote Originally Posted by ink883
    Yes, good call, keep in mind 807 was only after a hardreset and that went away soon after I restored my files. This is in line with the other thread about making a treo lightening fast (those users experienced only a temporary speed up).

    15% slowdown as the best case scenario isn't a good sign either. I certainly hope Palm didn't cheap out with slow memory modules.
    Actually, what's even stranger is that the OP got 844 on his "bone-stock" 700p, which is closer to a 10% difference from your stock 650:

    Quote Originally Posted by 2000 Man
    pbench data access = 1384% (only 844% on the 700p)

    Why is there ANY difference between the units?? Shouldn't they benchmark the same in stock form? Verizon vs. Sprint, maybe? That's another piece that needs to be charted. A QC issue, perhaps?
    Last edited by dcpmark; 07/06/2006 at 10:33 AM.
  10. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #190  
    Quote Originally Posted by ink883
    15% slowdown as the best case scenario isn't a good sign either. I certainly hope Palm didn't cheap out with slow memory modules.

    Quote Originally Posted by Rob_T
    Why couldn't they (Palm) take a 650, leave the button layout alone, add some RAM, set it up for an EVDO connection and otherwise leave it alone so they maintain the stability and other gains they already worked through on the 650?

    Perhaps it was intentional on Palm's part. I didn't have a 650, but from postings I've seen, many here are saying that the 700p is much more stable than the 650, even with all the updates. Fewer resets and crashes. As I've said, one theory is that the increased stability is the result of the memory changes.
  11. #191  
    Well, I seem to be on to something. I did a hard reset and my pBench data access jumped to 822. I then brought over USB Modem and VeriChat plus associated files, and pBench dropped to 800. I then synced my contacts and pBench dropped to 775. Then I brought over Chatter and all of its associated files and pBench dropped to 675.

    Yesterday, after deleting all my Java files, my pBench scores jumped 100 points.

    I am convinced that the pBench data access score goes down as you use up more storage space on the device, as my above experiment seems to show. I did a soft reset after installing each app and data files prior to running the pBench score. The most curious result of all was a loss of 25 points by simply syncing my contacts.

    I am therefore led to two conclusions: 1) The Treo 700p has a bug that results in lower data access performance the more storage you consume; or 2) The pBench results are erroneous since pBench was never designed to test the 700p's memory configuration. Regarding the latter, pBench may be testing the wrong parameter, or incorrectly calculating the result, due to the differences in how the 700p handles memory versus previous models. Remember, pBench has not been updated since 2003.

    I see no discernable difference between an 822 and 675 score as far as the use of my unit is concerned -- switching apps and screens is the same, and no freezes. I am intrigued that with essentially the same configuration as I had yesterday, my scores are 100 points lower.

    I think it's funny that we are judging our units based on a 3 year old benchmarking program that was designed for an entirely different CPU and memory architecture. How do we know that pBench is telling us anything meaningful, especially given the wild fluctuations in scores?

    My personal view is that I don't trust the pBench data access results. The program, being designed for totally different hardware and memory architecture, should not be trusted to benchmark today's equipment. Indeed, Palm made changes to the memory architecture between the 650 and 700p, so you cannot even compare the score of one to the other.

    Anyway, those are my findings. My overall experience with the Treo has been that it's a very solid, well performing PDA/phone. The incompatibilities and problems that I've personally experienced have all been due to 3rd party software issues.
  12. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #192  
    Quote Originally Posted by SonnyS
    A score of 1384 on pBench memory access is obviously from an overclocked or otherwise ROM-modified 650, and is therefore a dishonest number and will thankfully be disregarded by honest forum members.

    An honest result appears to be around 950, which shows less than a 20% difference from the 700p.
    Before he was banned, Chckhbrt posted a non-overclocked 650 score of 912. I just ran pBench on a friend's VZW 650 with very few third party apps installed: 885. I took a pic with the 700p, and will post it when I get home. His device showed 'VZW 1.01', so I don't know if he had updated his phone with all available updates. But the range now is as close as 844 for a 700p to 885 for a 650.
    Last edited by dcpmark; 07/06/2006 at 03:05 PM.
  13.    #193  
    For those of you who claim that pBench doesn't accurately test the database access times, please refer to Jeff Gibson's posts. He found the EXACT SAME PROBLEM on the 700p by measuring the time that it takes to access the database using the File Manager. It's no coincedence. pBench works and the 700p has data access problems.
  14. #194  
    Quote Originally Posted by 2000 Man
    For those of you who claim that pBench doesn't accurately test the database access times, please refer to Jeff Gibson's posts. He found the EXACT SAME PROBLEM on the 700p by measuring the time that it takes to access the database using the File Manager. It's no coincedence. pBench works and the 700p has data access problems.
    Say what you will, but I do not agree with your conclusions. You cannot judge a platform which runs on an entirely different CPU, a different version OS, and a new memory architecture than what the benchmark software was specifically designed for. It's just not possible -- you're comparing apples to figs.

    The fact that a difference between a 675 and 822 score on my machine makes no noticeable difference is additional proof in my book. However, if there is a problem that I'm just not seeing with my particular combination of hardware/software, I fully hope that Palm fixes this issue for everyone's benefit.

    Also, is it fair to say that the majority of folks with slowdown issues are Sprint models?
  15. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #195  
    Quote Originally Posted by SonnyS
    I see no discernable difference between an 822 and 675 score as far as the use of my unit is concerned -- switching apps and screens is the same, and no freezes. I am intrigued that with essentially the same configuration as I had yesterday, my scores are 100 points lower.
    I, too, have not noticed any difference in the speed of app switching or longer white screens since I got mine on 5/27, even though I keep adding apps. Maybe you and I just aren't perceptive enough.....
  16. dcpmark's Avatar
    Posts
    560 Posts
    Global Posts
    562 Global Posts
    #196  
    Quote Originally Posted by SonnyS
    Also, is it fair to say that the majority of folks with slowdown issues are Sprint models?
    Mine is Verizon, but I recall at least one other person with Verizon posting their speed complaints.

    Sonny, you didn't happen to check for pTunes skipping after a hard reset but before you restored your backup, did you?
  17. #197  
    Regarding the skowdown issue. I noticed today a considerable speed increase when using the Palm standard battery vs. the Seidio extended battery. Weird? Are you using the extended battery orthe standard? Try each and see if you can tell a difference. Just a thought. It could be a power issue.
  18. #198  
    Quote Originally Posted by Robert K
    Regarding the skowdown issue. I noticed today a considerable speed increase when using the Palm standard battery vs. the Seidio extended battery. Weird? Are you using the extended battery orthe standard? Try each and see if you can tell a difference. Just a thought. It could be a power issue.
    Where is the speed?...with the Standard or the Seido...I am suspecting the Seido, but I want to make sure. You said you noticed the speed increase when using the Palm standard battery vs. the Seido Extended battery, not pointing out which one has the speed.
    at&t iPhone3G
  19.    #199  
    Quote Originally Posted by dcpmark
    I, too, have not noticed any difference in the speed of app switching or longer white screens since I got mine on 5/27, even though I keep adding apps. Maybe you and I just aren't perceptive enough.....
    But your 700p started out slow. I'm not surprised that you don't notice it getting even slower after adding more programs into memory. Try playing pTunes in the background while browsing the web or loading a large app like Chatter. Skip-free on the 650 but skip-o-matic on the 700p. Maybe Jeff Gibson and myself are mistaken or maybe there really is a database access problem on the 700p. Hmmm.
  20. ink883's Avatar
    Posts
    872 Posts
    Global Posts
    883 Global Posts
    #200  
    Quote Originally Posted by Robert K
    Regarding the skowdown issue. I noticed today a considerable speed increase when using the Palm standard battery vs. the Seidio extended battery. Weird? Are you using the extended battery orthe standard? Try each and see if you can tell a difference. Just a thought. It could be a power issue.
    That is interesting, because the palm battery is 3.7v while the Seidio is 3.6v.

    Another possibility could be you noticed a speed increase because changing the battery would have resulted in a soft reset.
    Visor --> Visor Platinum --> Treo 300 --> Treo 600 --> Treo 650 --> Treo 700p --> Treo 755p --> Treo 800w --> Palm Pre
Page 10 of 16 FirstFirst ... 56789101112131415 ... LastLast

Posting Permissions