Measure your Notebook CPU Speed

My did it in 2min 30sec

hp :: Athlon XP 2600+ 2.0 :: ze4800 :: 512Mb :: 30Gb ::

 

<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by Scalded Ape

 

my results:
acer aspire 2012WLmi
P-M 1.5
512 mb ram
60 hd (4200)

00:02:25 for 2 million digits, curiously i had the same result for both battery and plugged in...any thoughts?

also, i ran it on my partners computer, acer tm 291, p-m 1.4, 512 ram, and the results were:
battery- 00:02:48
ac - 00:02:41

i'd be happy if more acer users jumped in to this really interstering thread

hope all's well,

 

<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by twister

 

Sony VAIO PCG-GR300P(CTO) Pentium III-M 1133MHz, 512MB RAM 40GB HD

AC 4:40
Battery, max battery life 5:49

Taskinfo tells me I'm only running at 1126MHz normally, and 729MHz in battery life mode. There's also "performance" mode which leaves the peak at 1126MHz, so the battery time was again 4:40 there.

The numbers don't seem right. At 729MHz that's 2/3rds of the processor speed on AC. That speed should have made it take 50% longer, but it only took 25% longer. I guess that could also mean that at full CPU speed the memory bandwidth isn't keeping up, otherwise on AC it should have only taken around 3:46.

Well. Soon I'll have a Dothan and these numbers won't matter to me any more...

 

<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by Andrew

Your RAM modules are not the same size. Dual-channel is not enabled in your computer as it is in mine. This only really affects this test, however. In most other scenarios the extra 256 megs of RAM would help more than the dual-channel capability would.
Don't fret[]
<hr height='1' noshade id='quote'></font id='quote'></blockquote id='quote'>

The Pentium M only works with the 855 chipset, which is a single-channel memory controller. So there is no dual-channel capability anyway.

 

Really, I had no idea, thanks for letting me know. I guess his is just messed up.

GPU cooling (100°C->75°C)* Inspiron 8600 * 1.8ghz Pentium M * 128 MB Radeon 9600 Pro Turbo (337/242 -> 400/300) * 2x256 MB DDR2700 SDRAM * Aquamark 3: 24058 * 3DMark'03: 3404 * 3DMark'01 SE: 13120

 

hey bootleg, would you mind compiling the data from fastest to slowest, I would like to compare these results to the PCMark results when they are in.

GPU cooling (100°C->75°C)* Inspiron 8600 * 1.8ghz Pentium M * 128 MB Radeon 9600 Pro Turbo (337/242 -> 400/300) * 2x256 MB DDR2700 SDRAM * Aquamark 3: 24058 * 3DMark'03: 3404 * 3DMark'01 SE: 13120

 

Toshiba 1135-S1553
2,4Ghz Celeron, 512MB DDR RAM (480 real, 'cause 32 is for video), 30 (28 real)GB HD 4200 RPM ATA 100, 64mb of video shared memory.

Windows partitions defined as: 10Gb for C: and 12Gb for E: (D: is DVDROM)

AC power: 3min 48s
Battery: 3min 55s

------------------------
Old one Toshiba 1135-S1553
Celeron 2.4
512 RAM DDR
64 Intel shared video
------------------------

 

Well, I reloaded the OS (Windows XP Pro), installed all my stuff back on there, and configured it how I like. Now when I run the program it does it in 1:55. That's a 15 second improvement, and the only thing different right now is that I made the C partition bigger on my hard drive.

I like to keep just the bare essentials on C to improve boot time and put the rest (music, movies, files) on a different partition that takes up the rest of the drive. So the first install had an 8 gig partition, which has been plenty in the past, but I didn't realize how huge PC games could be. So this time around I made it 25 gigs.

I never timed how long it took to boot before, but I don't notice a difference in it now. On the fresh install, it was booted up and ready to use 11 seconds after hitting the power button. After installing all my programs and stuff it takes a bit longer.

If anything, I have more stuff installed and more memory resident programs on it now than I did the first time I ran the calculation. The old time was 2:10.

Inspiron 8600, 1.8 P-M 745, 768MB 333mhz DDR RAM, 60GB 7200 RPM HDD

---------------------
i8600, 1.8Ghz Pentium M 745, 768MB RAM, 7200 rpm 60GB hard drive

 

Jetbook 9090a (which is a compal) Pentium 4 2.4 Ghz, 512 ram, 40 GB 4200, 64 MG non intergrated video. Firewall, Norton and wireless connection running in background

Plugged in 2 mins 47 sec

Battery 3 mins 8 sec

Re ran with nothing running in background

Plugged in 2 mins 44 sec

Battery 3 mins 7 seconds

Owner / Partner http://www.mtwolfcomputers.com

 

I found this forum looking for a P-M vs P4 comparison, and this site is great! Nice testing app. My desktop is an Athlon XP 2600+ (2.08GHz, 333FSB, 256KB L2), 512MB PC2700, 80GB 7200RPM HDD, FX5600 Ultra, and for the test, I got 2:00 flat

My grandmother's new Compaq R3000 notebook is over here as well. It has an Athlon 64 3000+ (1.8GHz), 512MB RAM, 60GB 4200RPM HDD, MX420 32MB, and for the test, it got 1:59 while up loading pictures on a wireless network on AC power. I will test it again after it's done with that.


Edit:

After the pictures finished loading, I restarted the Notebook and tried it again and received 1:57 on AC.

On the battery, it automaticly downlocks it to 1.6GHz. On that, I received 2:07
In Power saving mode, it clocks it at 800MHz. On that, I received 3:52

 

AC 2min 9 sec
Battery 2min 9 sec

Compaq R3000T (CTO)
P4 Desktop 3.0GHz w/HT
1 X 512MB RAM
Radeon 9600 128MB
60GB 5,400 RPM HD
No mods/overclocking: GPU: 310MHz, Memory: 190MHz
AquaMark: 19,796
PCMark04: 3955
3DMark03: 2508
Over clocked GPU: 378MHz, Memory: 256.5
Aquamark3: 24,257

 

Acer 2012WLMi
1.5 Ghz Pentium-M (Banias)
1gb ram
60gb 4200rpm hd

2:35 with quite a few things running: 3 i.e windows, one folder, one word 2003 document, mirc, bit spirit, nero drive speed, popup stopper.

I'm compiling a review of the laptop just now, so I will be doing a fresh XP Pro install (currently have XP Home) and do some fresh benchmarks.

- Odie

 

I was curious to find out the performance of my 2-year old desktop.

No Brand Name
Intel Celeron 1.3 GHz
256 MB
40GB HD at 4200rpm
XP Pro
AVG Anti-virus auto-protect running in background

AC Powered : 8min 5sec

 

My brand-new Acer Aspire 2012WLCi
1.5 GHz Intel Pentium M
512 MB
40GB HD at 4200rpm
Intel Extreme Graphics 2
XP Home
Still factory configurations + Norton Antivirus 2004

With AC : 2min 17sec
With Battery : 2min 15sec [:0]

With battery power it's 2sec faster. I ran both tests one after the other, after pulling out the AC jack. Battery fully charged. No power saving mode. Anyone got a clue ?

PS: Later I figured out that my notebook is running on M 715 Dothan processor. Several other Super-pi tests showed a stable value of 2:15 It doesn't seem to change with number of applications running or number of background services.

 

Toshiba Satellite M30-832
Pentium M 1.5GHz ; 512MB ; 40Gig H/D

2min 21sec = Plugged in
2min 21sec = Battery(Normal Mode)
3min 56sec = Battery(Long-life Mode)

Running XP-home & 46 processes at present, plus 6 web pages(over Wi-Fi & Sygate firewall)



--- --- --- --- --- --- ---
Keep young ......
Stay immature !!
--- --- --- --- --- --- ---

 

Desktop: Athlon 2600 2m 12
Notebook: Compaq Presario 2199us 2800 Athlon M 40GB 7200
2m 19 Plugged
4m 07 Unplugged

 

Dell Inspiron 8600 M1.7 512DDR333, 80gb 7200rpm, radeon9600 pro

AC 1.51
Battery 2.01

Medion P4 2.5, 512DDR266, 40GB, GF4 MX440 Go

AC 3.14
Battery 4.41

 

Asus M6Ne Dothan 2.0 GHz, 2GB RAM, 60GB 7200rpm
AC 1:52, max battery life 4:21

It's definitely a great improvement over my old VAIO, but it's no faster than some of the 1.7 and 1.8GHz Pentium Ms reported here. Either there's something wrong on my system configuration, or this test saturates the memory bus bandwidth thus making CPU speed irrelevant.

 

hey guys,

2:10 AC power
3:12 Battery

Dell Inspiron 5150 3.2GHZ, 512MB PC2700 ram, 80GB HD (?RPM), geforce FX5200go O/C 241/510 8247 3dmarks (2001)

 

* HP Pavilion zv5000z * AMD 64 3000+ 1.8ghz * 256mb * 30gb 4200rpm *

AC POWER
Time: 1 min. 53 sec.

[8D]

 

E-machine m6811 laptop,
Mobile AMD 64 3400+
512Mb Ram
80GB 4200 HD
ATI Radeon 9600 64MB

AC: 1:40

 

i just got this notebook, and so i decided to see, for fun, what it could really handle. it's running now (as i'm typing) with firefox (obviously), winamp with 300 songs loaded (playing too), two instances of aim, norton antivirus and liveupdate, spybot, and varioius dell system maintenance programs. all on max battery.

here are the computer specs (inspiron 600m):
pentium m 755 2.0ghz
14.1 sxga+ (brightest setting)
256mb ddr266
32mb ati mobility 9000

max battery with a load of programs running: 5:35 (but no slowdown at least!)

 

i then ran it with nothing running in the background, on max performance with ac connection and got: 1:58.

shouldn't that be faster? people with 1.7 and 1.8 dothans have reported lower times. can someone tell me what's wrong?

 

I ran after a clean reboot (only on sec difference from when i was running a bunch of programs the first time).

presario 2190us
2.4g celeron (256k L2 cache), 512m memory (-32 video mem), no swap, 5.4k IBM/hitachi HD 8m buffer (gets pretty hot)
========
5:19

I'm waiting on my zt3000 that i ordered last week, it's a 1.7 dothan so i'm sure it will do a ton better then this celery crap.

IMHO the best desktop replacement is anything with a >=3g Athlon 64
the best mobile notebook is a >=1.6 dothan

If I didn't travel so much i would have certainly prefered a 3G Athlon 64 chip.

 

Flockey:
In order to get your question answered faster, you should just start a new topic.

*******************************************************
Fujitsu S6210: 1.6Ghz PM ~ 768MB RAM ~ 60GB 7200RPM Hard Drive
*******************************************************

 

you're right it looks like i'll have to.

 

2:12 AC Power

Compaq R3140CA
AMD 64 3000+
512mb DDR PC2700 Ram
60GB HDD 4200RPM
15"4 WXGA TFT
GeForce 4 420MX GO 32 MB

 

since i have no laptop (yet), i'll post my desktop score.

AMD Athlon XP 1700+ (1466 GHz)
512 MB PC2100

+ 000h 00m 00s [ 16K]
+ 000h 00m 01s [ 32K]
+ 000h 00m 02s [ 64K]
+ 000h 00m 06s [ 128K]
+ 000h 00m 15s [ 256K]
+ 000h 00m 36s [ 512K]
+ 000h 01m 20s [ 1M]
+ 000h 02m 59s [ 2M]
+ 000h 06m 33s [ 4M]
+ 000h 14m 24s [ 8M]
+ 000h 34m 43s [ 16M]
+ 001h 13m 39s [ 32M]

i'll post my laptop scores as soon as i decide what to buy and then buy it

CJ

 

CSpwntsYOU: I would have guessed yours to be faster. What did you have running in the background?

*******************************************************
Fujitsu S6210: 1.6Ghz PM ~ 768MB RAM ~ 60GB 7200RPM Hard Drive
*******************************************************

 

2:10 battery and 2 on AC

Inspiron 9100- 3.2ghz P4 HT - 1gb RAM - 100GB HD - 4x dvd/cd burner -128mb ATI 9700. RoX0r mY coX0r!

 

Yeh... you think with all that ram would significantly decrease the calculation time.

 

Pentium 4 2.8 HT, 512MB RAM, 40 GB 5400rpm

AC power: 1:52

 

nothing, i dont know why the numbers are the way they are, would having alot of videos saved slow it down?

Inspiron 9100- 3.2ghz P4 HT - 1gb RAM - 100GB HD - 4x dvd/cd burner -128mb ATI 9700. Too Much Computer for Such a n00b!

 

i dont know, but dont worry about the results, as this benchmark doesnt mean a whole lot.

*******************************************************
Fujitsu S6210: 1.6Ghz PM ~ 768MB RAM ~ 60GB 7200RPM Hard Drive
*******************************************************

 

Hi run1track,
This test means just as much as any other for the purpose for which it is being used; to gauge CPU mathmatical processing performance. The programs results are nearly entirly based on CPU mathmatical processing horsepower, the next most important component is the memory speed/bandwidth. This is where CSpwntsYOU may want to look if he does not have much running in the background (more than 2% of CPU is being used for background tasks) as Dell is notorious for using slower memory in their notebooks to shave costs.

Thanks for participating in this everyone
Jack

"They who would give up an essential liberty for temporary security, deserve neither liberty or security" (Ben Franklin)
http://pbase.com/joneill

 

I have brought the issue b4. The topic name is inappropriate, it should have been, "Let's measure our notebooks' CPU speed"
<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by Run1track

 

Dell Inspiron 2100 Pentium III 700Mhz, 256 MB SDRAM PC100

Running IE, Excel, and some other Dell "stuff"...6m51s on AC woohoo!!! [ ]

 

Thanks, Freeman.

<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by freeman

 

Took 1 min, 53 secs on my emachines M6811 plugged in and 3 min, 59 secs on battery power.

 

With a degree in computer engineering (among several) I find some of the comments on this list quite wrongheaded. The super_pi benchmark is neither a useful comparison of overall laptop performance (no one said it was) nor a pointless exercise (as several have claimed). Personally I think it says something useful about one specific grouping of system performance paramters and even more about common misperceptions regarding system performance issues.

Q: If I add more memory shouldn't it run faster ?
A: No ! The Windows version of this benchmark will occupy about 17MB of SDRAM and Mr.Gates operating system (any versions) will reserve an unconcionable amount of memory. Beyond the 17MB and the system IO buffers more memory is of no help at all. This question is roughly like asking if you add more tires to your car will it go faster .... the answer is generally NO unless you have a flat (or are have a process that is paging). The silly notion that more memory is a cure for processing ills is largely due to the uncontrolled use of memory by badly written code.

Q: I have a faster processor, shouldn't the processing time be shorter time ?
A: Generally yes, but architecural differences have make up for many megahertz of clock speed. This particular bechmark uses the processor intensively, the CPU-memory interface intensively and the disk IO system somewhat. The 'Help' and text on the original ftp-site claim the task is disk IO intensive, however that is not the case for modern systems ... The IO times are well under 10% of the total process time.

Q: I have a dual processor (or hyperthreading) so it will run haster - right ?
A: No. This is a variation on the "adding more tires" fallacy. This particular process *could* have been written as a multi-threaded process or a group of co-operating processes, in which case it could take advantage of HT or multi-cpu. It is currently written as a single thread so there is no advantage (and a potential disadvantage) to multi-cpu systems.

Let me briefly explain the potential disadvantage of multi-CPU systems with a single threaded process. Typically each CPU will have a separate L1 cache, perhaps share an external L2 cache or have separate L2, and certainly share SDRAM. When a thread is running it uses the L1 and L2 caches which partly fill with the most recently used data and instructions for that process. These caches can be of considerable size ... megabytes. Periodically and on events (IO interrupts etc), the operating system will "reschedule" the thread that is running in a processor. Perhaps every few milliseconds the OS will re-evaluate whether that thread should be allowed to continue or else be replaced with a more compute-worthy thread. The OS re-eval is cheap, *but* on some OSes the thread under scrutiny may be re-assigned to the other processor(or HT engine) for no good reason. After all this is "SYMMETRIC" multiprocessing and in many ways the OS makes no distinction at all between the different processing engines. Unfortunately the act of restarting the thread on an alternatate engine requires that all unshared caches be flushed to SDRAM and then as the thread continues, instruction and memory accesses on the new engine will be met with cache-misses which slow the access to SDRAM speed. This can have a huge impact on performance.

In any case the Windows "Task Manager" can provide a great deal of insight into how such a process runs.

I'm searching for a new laptop, but I've run SUPER_PI on several desktop platforms all have 2.6Ghz P4 w/HT and reasonable fast (PC3200) memory systems and varying disk systems.

Homemade ASUS P4C800E P4 w/HT 2.6Ghz:
Win2K, no optimization: 173 seconds (2min 53sec)
Win2K, set process affinity to CPU1: 123sec (2min 3sec)
Win2K, safe mode, defrag disk, affinity to CPU1: 122sec (2min 2sec)

Homemade MSI NEO-FIS2R P4 w/HT 2.6Ghz:
Linux (RH9): 97 seconds (1min 37sec)
Linux, SDRAM disk:Win2K, set process affinity to CPU1: 123sec (2min 3sec)
Win2K, safe mode, affinity to CPU1: 122sec (2min 2sec)

 

Sorry - the last few lines above are wrong I <tab> entered prematurely.

HomeMAde MSI Neo-FIS2R 2.6Ghz P4 w/HT:
Linux (RH9): 97 sec
Linux (RH9), sdram disk: 95 sec

Only a 2 second improvement when using SDRAM in place of a disk.
Also the Linux process only used about 1MB of memory vs 17MB for the Windows version.

Note that most of the 25% improvement from Win2K to Linux (122sec -> 97sec) is likely due to the compiler technology. Gnu compilers rock ... MS tools not so much. I wish I had source code so I could recompile this benchmark for Windows with the GNU compiler. I'm no pro-Linux bigot, but it's hard to avoid the conclusion that MS has lousy toolchains which squander cpu performance and memory. Thanks to these problems from Redmond your 3Ghz & 1GB whatzis is really running more like a properly running 2.2Ghz 256MB system.

I don't care terribly that the scheduler on Win2k can't make a thread 'sticky' to a particular processing engine to save 40%(173sec->123sec). This is certainly not the case for Linux and is reportedly fixed in WinXP, and it doesn't represent a real-world problem in systems with multiple competing processes - the way I normally operate. It does harm single thread benchmark performance and also single thread apps (perhaps including some games) - so beware.

Steve

 

Oh yeah ....
calculating the max 33million places of pi took
44 min 37 sec (2677 sec) on the Linux system above using
a plain-jane 80GB WD EIDA drive.

 

Don't tell me you are running top in another shell? cuz, that CPU resource could have gone into SuperPi[ ] Not that it would make much difference, or any useful.
<blockquote id='quote'> quote:<hr height='1' noshade id='quote'>Originally posted by stevea

 

AC plugged: 3:59
Ac unplugged: 5:08

Anh Nguyen.

Notebook: Sony PCG-X505P, 1394 DVD-RW (PCGA-DVRW1).
Desktop: D865GBF, 2.6G-HT, 2x256M, 2x73.4G Raid 0, 18.1" Sony S81R/B LCD, A2ZS Pro, DVD-ROM.

 

well stevea seems so anxious with superpi there...
haven't tested it with 2million..but i really wonder why it was started with using 2million?any particular reason?why not 1 instead of 2? the result will be way faster...
for 1 million my result was 45 seconds..
my amd [email protected] was 41 seconds..
not bad, isnt't it?

 

I agree that one million would be nice.

*******************************************************
Fujitsu S6210: 1.6Ghz PM ~ 768MB RAM ~ 60GB 7200RPM Hard Drive
*******************************************************

 

Hmm... Obviously the results are only useful if everyone uses the same (Windows) binary, but I just ran the Linux version *in my Linux VMware partition under Windows* and it still ran faster, 1:37. That tells me there's too much difference between the Linux and Windows binaries to really compare them, but it's interesting nonetheless.

old: Sony PCG-GR300P 1.13GHz PIII-M, 512MB
new: Asus M6Ne 2.00GHz P-M, 2GB

 

Well stevea, you're not the only computer engineer hanging out here. I'm still a bit perplexed at why my 2GHz Pentium M is yielding identical run times to the 1.7 and 1.8GHz numbers posted so far. There ought to be about a 10% difference, around 10 seconds or so, if this is a pure CPU speed test. My only explanation is that this test is memory-bound. I haven't run it on Linux yet; I'm surprised that the Linux version only uses a 1MB buffer to calculate the same 2M digits. Anyway, since the program allocates a 16MB buffer on Windows, one can only assume that its working set exceeds the 2MB L2 cache of the Dothan chip.

Setting the process to High priority and then exiting TaskInfo got my best time down from 1:52 to 1:51, so there's obviously not a lot of overhead wasted in having a monitoring program running.

Other benchmarks (like Sisoft Sandra) show a slight improvement in memory bandwidth by running a lower screen resolution, but that had no measurable impact on my SuperPi result.

Since I'm using DDR333 CL3 memory, I'm curious about whether some of those faster times were reported using DDR400 memory.


old: Sony PCG-GR300P 1.13GHz PIII-M, 512MB
new: Asus M6Ne 2.00GHz P-M, 2GB