What's up with the SLOW Dual Core CPUs?!

[underclocker]

I've noticed that while some CPU's are getting faster, like the 2.4GHz dual cores with 4MB caches, there are some very low offerings on the market, too.

For instance, Lenovo sells R61's with the T5250 CPU, which is only 1.5GHz. They sell lower CPU'd machines, as well.

Are these manufacturers just ripping off unsuspecting consumers? Are these machines any faster than Pentium M based machines? Are they slower?!?!

[qviri]

The Core architecture is significantly faster clock-for-clock than Pentium Ms. The T5250 (which is a C2D, even faster than CD) may not be as fast in single-core applications as a 2.26 GHz Dothan, but it's definitely faster than a P-M clocked around 1.6 GHz, and the second core helps immensely in some applications.

[mattbiernat]

I've noticed that while some CPU's are getting faster, like the 2.4GHz dual cores with 4MB caches, there are some very low offerings on the market, too.

For instance, Lenovo sells R61's with the T5250 CPU, which is only 1.5GHz. They sell lower CPU'd machines, as well.

and lenovo is really smart. i mean what are most computers used for? I bet you #1 is internet #2 media #3 MS office. Most people don't go behind number three. For most there is no need to have the newest and fastest CPU.

[underclocker]

I completely agree that most people will use their machines for non-CPU intensive tasks, like surfing and Quicken, but I do think that for most computer buyers, the current offerings are a little tricky.

I think that many people assume when they are buying a new computer after 2 to 4 years that the new model will be significantly improved. It seems that someone moving away from a three year old 2.0GHz R52 for a new dual-core R61 might be shocked that there's little difference or possibly a reduction in performance!

I'd love to see some comparison numbers of the most recent Intel offerings to similarly clocked machines from a few years ago.

[aaa]

I completely agree that most people will use their machines for non-CPU intensive tasks, like surfing and Quicken, but I do think that for most computer buyers, the current offerings are a little tricky.

I think that many people assume when they are buying a new computer after 2 to 4 years that the new model will be significantly improved. It seems that someone moving away from a three year old 2.0GHz R52 for a new dual-core R61 might be shocked that there's little difference or possibly a reduction in performance!

I'd love to see some comparison numbers of the most recent Intel offerings to similarly clocked machines from a few years ago.

Thing is, the people who will notice the performance difference are also most likely to notice to begin with; "Wow, this is only 1.5ghz. Maybe I should upgrade to 1.8 or 2.0."

Although, yes, 1.5ghz Core 2 is a bit of a downgrade compared to a 2ghz PM, though obviously not as big as the numbers imply (probably only 10%).

[mattbiernat]

Although, yes, 1.5ghz Core 2 is a bit of a downgrade compared to a 2ghz PM, though obviously not as big as the numbers imply (probably only 10%).

yeah but 1.5ghz core 2 will do much better executing several applications at once compared to 2ghz pm. so is it really a downgrade? more like a trade off i think...

[aaa]

Although, yes, 1.5ghz Core 2 is a bit of a downgrade compared to a 2ghz PM, though obviously not as big as the numbers imply (probably only 10%).

yeah but 1.5ghz core 2 will do much better executing several applications at once compared to 2ghz pm. so is it really a downgrade? more like a trade off i think...

For some reason I've never noticed the benefits of dualcore in everyday tasks, in fact the only time I find it useful is the rare occasion of having an app peg the cpu at 100%, and still be able to have other things behave just as snappily.

However, when using one of the very prevalent single-threaded applications (like Firefox), I can feel just how many megahertz a cpu has, and dualcore doesn't help much in the single-threaded department.

Top that off with Core 2 being only 5-10% faster per mhz doing normal things (as opposed to 20% faster when encoding video or similar). I would never go from a 2ghz single core to a 1.5ghz dual, as I would feel the difference.

[dr_st]

At any given point there's low-end, mid-end, and high-end.

Even though the general trend is newer=faster, older high-end and newer low-end can definitely overlap.

I assure you that the T5250 @1.5GHz, which is the low-end of R61, will easily beat the low-end of R51 (which was what? A Celeron M @1.3GHz?)

Just like a GeForce 8800 is faster than a GeForce 7800, but the 7800 is faster than a, say, 8400.

If you had a top-of-the-line T4x, which cost $2000, and you are replacing it with a bottom-end T6x for $1000, why would it suprise you that you're actually getting less performance?

[ulrich.von.lich]

I'm kinda confused with the Core Duo thing too. They are generally a lot faster because we often run many applications (actively or in the background) simultaneously nowadays.

The Santa Rosa platform has the dynamic acceleration technology which turns off one of the CPU cores and overclock the active one while doing single threaded applications. I assume that would make the 1.5Ghz higher. Obviously the frenquency is not the only thing that decides the CPU's speed. i.e. the PM 760 2.0Ghz is faster than PM 765 2.1Ghz while the latter has the higher frenquency.

I often find my CPU is 100% occupied. For example, today I was having lunch, burning discs, watching family guy and scanning documents (my scanner is very slow so I have to find something to do at the same time.) I already find the PM 760 kinda slow to handle them competently.

[wearetheborg]

I'd love to see some comparison numbers of the most recent Intel offerings to similarly clocked machines from a few years ago.

A core 2 duo 1.5Ghz will be faster than a dothan 2ghz.
see
http://forum.notebookreview.com/showpos ... count=1053
(and that is for core duo, C2D will be faster).

[aaa]

I'd love to see some comparison numbers of the most recent Intel offerings to similarly clocked machines from a few years ago.

A core 2 duo 1.5Ghz will be faster than a dothan 2ghz.
see
http://forum.notebookreview.com/showpos ... count=1053
(and that is for core duo, C2D will be faster).

I don't think super_pi really applies to things in general (that is what was used, right?).

[wearetheborg]

Yes, super_pi was used.

[SHoTTa35]

Remind me of the P4 days. Remember those 1.8Ghz P4s CPUs that felt slower than the 1.3Ghz P3 they replaced? I remember i was helping a lady setup her brand new Dell Dimension system. It was all new P4 and shiny. I had a Dell Latitude C610 with a 1.2Ghz P3 and my system ran apps faster than hers. She had like tons of RAM and faster hard drive and all.

That's the reason also that the Pentium M when they came out was running neck and neck with a 2.2Ghz P4 while the PM was at 1.7Ghz or even 1.6Ghz. So lower clock saving power and they were still kicking the HOT and HUNGRY Pentium-4Ms.

Clock speed isn't all that big of a deal anymore which is why CPUs stopped heading for the 3Ghz barrier. Each new CPU comes out starts out low again and pushes up to 2.6Ghz then a new one comes again.

[kosse]

Remind me of the P4 days. Remember those 1.8Ghz P4s CPUs that felt slower than the 1.3Ghz P3 they replaced? I remember i was helping a lady setup her brand new Dell Dimension system. It was all new P4 and shiny. I had a Dell Latitude C610 with a 1.2Ghz P3 and my system ran apps faster than hers. She had like tons of RAM and faster hard drive and all.

2x. The early Pentium 4 machines were a terrible disappointment. I bought a then new 1,8Ghz P4 laptop in 2002 and it didn't run games much better than 1GHz P3 machine with similar GPU. Willamette P4 had bad performance to start with and SDR SDRAM in laptops only crippled it more. To add to the misery it got very hot and the small annoying fans were running at top speed all the time.

[Thinkpaddict]

Yeah, the P4 line was a disappointment. Intel dropped the ball on that one, it is almost as if they got caught in the Ghz race with AMD and just focused on raw clock frequency. Luckily the Pentium M continued where the Pentium III left off.

I bet you that the Intel Core Duo is much better than what it seems to you with regards to the Pentium M. There are a lot of factors to consider when you expect an increase in performance from a computer. If your limiting factor in performance was I/O (Hard drive, RAM), then there is nothing that a better CPU can do to help you.
Take a look at the CPU usage meter while you perform some tasks, and I am sure it will be well below 100% almost all of the time.

As far as using the double core architecture to its full extent, if you have a multitasking OS (which one isn't these days?), then the only time it would be beneficial to run both cores simultaneously is if one of them is already maxed out. I don't know how Windows manages dual core architectures, but it would be interesting to see some comparative measures between different operating systems and different task loads.

[K. Eng]

I think that there is no need to put super fast CPUs in most notebook computers. Notebooks benefit more from lots of RAM and faster HDDs.

Also, bloatware has been destroying much of the gains in computer speed. Bootup speed on a lot of consumer grade machines is abyssmal, and RAM gets filled up with useless processes, forcing Windows to use the page file.

[leoblob]

...Clock speed isn't all that big of a deal anymore which is why CPUs stopped heading for the 3Ghz barrier. Each new CPU comes out starts out low again and pushes up to 2.6Ghz then a new one comes again.

Perhaps the yield rates are so low at 3+GHz that it's actually cheaper for them to make multi-core processors running at slower clock speeds?

I wonder if people have been duped into thinking these multi-core processors are great (even though many apps are still single-thread) simply because the processors are too hard to fabricate at the high clock speeds?

If you were the average computer user, I suspect you'd find a "5GHz P4" much more appealing than a multi-core 2 GHz processor... if only they could manufacture them ... ??

Just speculating...

[qviri]

If you were the average computer user, I suspect you'd find a "5GHz P4" much more appealing than a multi-core 2 GHz processor...

And neither processor would be faster in real life than a single-core non-P4 2 GHz one because your computer is still held down by an ancient storage subsystem which offers speeds measured in megabytes per second...

[Thinkpaddict]

If you were the average computer user, I suspect you'd find a "5GHz P4" much more appealing than a multi-core 2 GHz processor...

And neither processor would be faster in real life than a single-core non-P4 2 GHz one because your computer is still held down by an ancient storage subsystem which offers speeds measured in megabytes per second...

That's a very good point. I/O bandwidth is maxed out more often than CPU capacity.

Regarding dual core systems being able to benefit only with multithreaded applications, I think that's just not true. You run more than a single process at any time (I don't even mean just end user apps with GUIs that you see, but also any process, including OS kernel, modules, services, etc.) It should be possible to allocate different processes to each core, even if the processes themselves are not multithreaded. The point is that unless one of the cores is maxed out, I don't think there is real benefit in using the other one (although I have the feeling this is wrong, someone please correct me!)

[qviri]

[contents of this post deleted by author]

[Troels]

(snip) The T5250 (which is a C2D, even faster than CD) (snip).

Actually, it is far from. The XX50 is a 533 MHz FSB offering, while the XX00 series (from CD and C2D) are 667 MHz. RAM speed is at 1:1, while the Santa Rosa C2Ds is 800:667.

yeah but 1.5ghz core 2 will do much better executing several applications at once compared to 2ghz pm. so is it really a downgrade? more like a trade off i think...

The reason why two cores doesn't do double the performance of an equal single core, same architechture CPU is the programs that needs to be executed. Writing programs for multicore processors that distributes processes equally to both processors is not always made very sucessfully.
Many larger programs do not even support multiple cores.

[leoblob]

Unfortunately, this discussion is starting to go over my head.

FWIW, I will offer up one observation... When I run Firefox on my Intellistation (with two separate Xeon processors) and I'm doing something intensive like playing back a youtube video at 1280 x 1024, a significant amount of load is handled by each processor. It's usually not too well balanced (like 25% / 75%) but it's definitely split up to a degree.

As I tried to articulate before, I noticed years ago that clock speeds never seemed to get past around 3.4GHz... now I almost feel that these multi-core processors are a "work around" for being unable to manufacture high clock rate processors.

By contrast, I suspect that it's pretty easy to make processors with today's "low" clock speeds of around 2GHz, so the primary issue with multi-core processors might simply be heat generation/dissipation... plus whether or not the software makes optinum use of the multiple cores.

[qviri]

[contents of this post deleted by author]

[underclocker]

Excellent conversation so far. Very interesting. I'd agree that computers today are generally faster and better vaues that computers two years ago. However, some of the credit for this goes to components other than the CPU. The hard disks, memory (bus), and video are faster.

Let me add another dimension to my original questions. Intel offers dozens of different CPU's for laptops - all at the same time, there are dozens of current CPU's available for laptops!

There are also seven CPU classes; Celeron, Celeron M, Core Solo, Core Duo, Core Solo 2, Core 2 Duo, and Pentium Dual Core.

And the model numbers are very similar and nearly impossible to determine which falls into which class, based on the number. For instance, a T2330 is a Pentium Dual Core and a T2300 is a Core Duo.

Additionally, a higher number, doesn't necessarily mean a faster speed.

I think Intel is committing "product obfuscation". They are trying confuse the public by flooding it with competing and overlapping products. They've perfected this marketing techique (there could be another name for it - it's almost like bait and switch.) They introduce a worthy (fast, advanced) CPU, get good press, then start to splinter it and ride the success that the introduction earned.

Is there a better explanation for the Core 2 Duo T5270 (1.4GHz, 2MB cache, 800MHz FSB)? And which is faster the T5270 or a Pentium Dual Core T2130 (1.86GHz, 1MB cache, 533MHz FSB) or a Celeron 450 (2.0GHz, 1MB cache, 533MHz FSB) or a Core 2 Duo T5300 (1.73GHz, 2MB cache, 533MHz FSB) or a Core Duo T2300 (1.66GHz, 2MB, 667MHz FSB)?

Product Obfuscation!

[leoblob]

...And the model numbers are very similar and nearly impossible to determine which falls into which class, based on the number. For instance, a T2330 is a Pentium Dual Core and a T2300 is a Core Duo.

Additionally, a higher number, doesn't necessarily mean a faster speed.

I think Intel is committing "product obfuscation". They are trying confuse the public by flooding it with competing and overlapping products. They've perfected this marketing techique (there could be another name for it - it's almost like bait and switch.) They introduce a worthy (fast, advanced) CPU, get good press, then start to splinter it and ride the success that the introduction earned.

Is there a better explanation for the Core 2 Duo T5270 (1.4GHz, 2MB cache, 800MHz FSB)? And which is faster the T5270 or a Pentium Dual Core T2130 (1.86GHz, 1MB cache, 533MHz FSB) or a Celeron 550 (2.0GHz, 1MB cache, 533MHz FSB) or a Core 2 Duo T5300 (1.73GHz, 2MB cache, 533MHz FSB) or a Core Duo T2300 (1.66GHz, 2MB, 667MHz FSB)?

Product Obfuscation!

Thanks. I just knew we were getting screwed somehow.

[mattbiernat]

Is there a better explanation for the Core 2 Duo T5270 (1.4GHz, 2MB cache, 800MHz FSB)? And which is faster the T5270 or a Pentium Dual Core T2130 (1.86GHz, 1MB cache, 533MHz FSB) or a Celeron 550 (2.0GHz, 1MB cache, 533MHz FSB) or a Core 2 Duo T5300 (1.73GHz, 2MB cache, 533MHz FSB) or a Core Duo T2300 (1.66GHz, 2MB, 667MHz FSB)?

Product Obfuscation!

is there really such a big difference between any of these? Okay... 1.4Ghz versus 2.0 Ghz yes, but 1.73, 1.66 and 2.0 Ghz? Common difference is going to be minimal. Perhaps 10-15%. 1MB cache versus 2MB cache isn't also going to make much difference unless you are doing a lot of gaming and other intensive applications. I agree thou, there seems to be something strange going on here. Why offer 1.4 Ghz, 2MB and 800 FSB and 2.0 Ghz 533 FSB and 1MB cache? But if you consider that the first choice was the top product of 2005 while the latter was economy of 2007 it all makes sense.

[Truthfinder]

The large amount of ram that is used in these newer units is what really make the difference.

Ram is key, not so much weather or not you use a 1.8Ghz or a 2.0 ghz processor.