• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Celeron M

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Sure mobile Barton will OC to 2.7 GHz, but that's not what I'm asking here.

It's also pretty obvious cache affects real world performance, but I'm after how much impact it has. If you recall, the difference between a same clocked barton and thorton is worth around 100 MHz, which is like 5% if the clock speed is 2GHz. And we all know how Celerons are way more than a 5% difference from the P4. But my question, is the difference between P-M and Cel-M only as much as that between a P3 and Tualatin Celly?

I'm thinking, if I were to actually get a laptop, will Celeron M offer a good performance for the buck?

Afterall, Pentium M is based on P3, as is the Celeron M, so I'm suspecting a Celeron M will have a higher number of instructions processed per cycle than a P4. And tualatin celly's did pretty well against the P3s.

If that's the case, Celeron M would be a better choice than regular P4 Celerons for laptops. But I still would need to compare with the mobile XP2500.

Something I generally dislike about hardware review sites is that they rarely provide information on the low end of hardware. You can easily find loads of information on Radeon 9600, 9800XT and so forth, but if you want benchmarks for a Radeon 9200SE, you aren't going to find it at major sites.

Sure low end product reviews aren't going to help the companies make a load of money like they do with the mid and high end stuff, but for those with less deep pockets, information on low end stuff is useful.
 
The Gateway M505X

Hey Guys I need some help. I have a The Gateway M505X notebook Intel Pentium M processor 1.4GHz Impressive 800MHz front-side bus for increased system throughput
ATI Radeon m10 9600 graphics with 64MB DDR memory . I like to oc my M processor 1.4GHz too see hy much and can get it too :) I need a oc programs.
 
One thing you need to remember about the Banais (sp?) core. IT CANT SCALE

It has been like 6-months to a year since the 1.6ghz and the launch of the platform and the only speed bump was 100mhz to 1.7 ghz. Very Very Very Bad.

Look the coppermine p3 had a problem hitting 1 ghz. And the Banais is based off of that (actually based off of tualatin). And when you think that the core came from the Pentium Pro that couldn't ever go past 200 mhz, you see the P6 core has been really pushed to the breaking point. From 150mhz to 1700 mhz. YOu do need to pat intel on the back for a core that can go that far. Maybe a 90nm die shrink will get us to 2.3 ghz but thats the end I think.
 
I thought Intel just made an announcement to move Banias to desktop :rolleyes:

Slow and efficient can be as good as fast and less efficient. AMD Athlon64's clock speed won't hit 3.4 GHz like P4s, but the performance will match it.

If Banias is going to run 1.7 GHz, that's fine by more, although it maybe a drawback for people who just want more GHz regardless of efficiency.
 
barton2500 said:
tualatin celly's did pretty well against the P3s.

That's because most desktop P3s had 256k L2 cache and Tualatin celerons also had 256k L2 cache. No cache difference to compare there, although I think the cache latency was increased on the celeron.

And remember, there have been changes from the P6 arch in Banais. What is the Cache difference between the Pentium M and the Celeron M? I think that as caches getlarger and larger, the performance increase form increasing cache size will decrease.
 
1. is there a mobile duron?!?!?! i dont think there is and i dont think AMD wud waste money on making mobile duron wen they got the athlon with is near anough as cheap as the duron for budget people
2. if there was a mobile duron, it wud kick celeron butt :p
 
Mobile duron's max speed is 1.3 ghz and is morgan core based I think, it is for sure a ceramic processor which means it is very old. There is no applebred mobile duron.
 
The Pentium M and the Celeron M are not simply "Pentium Pro architecture pushed to the limit"... They have the same basic instruction and ALU core of a P3 (less stages, more work per clock cycle) but also have the bus interface of a P4 (quad pumped 64-bit cycles). Combined with much more advanced power management features and the rest of the Centrino bus options, there is also a considerable amount of reworking involved. This isn't something that was just overclocked to hell and made mobile.

The result is a chip that performs >30% better than a P4 because of it's much higher IPC and equal memory bandwidth, while sucking less power than a much lower clocked P3 because of the smaller micron process and the more advanced power and clockspeed modes.

If the only difference between the Pentium M and the Celeron M is the cache size, then it's likely that a Celeron M of 2.0ghz will still outperform a standard P4 at the same speed. The reason? Normal P4's require a lot of cache because of their very long pipelines: if a stall happens and the data needs to be fully refreshed, they need a lot of cache to hopefully retain all that data. Hence the reason why a P4 Celeron with only 128kb of cache is such a terrible performer: throw it a few branch mispredictions in that long-assed pipeline and the performance plummets because the cache is likely already housing different data.

But the Celeron M (again, so long as it's the same core as the normal Pentium M) isn't going to have such a massive branch mispredict penalty because the pipelines are shorter. That, and it still has 512kb of cache to go back on. You basically have the branch mispredict-recovery of a normal P4, while being significantly more efficient in it's processing abilities.

Hope all that babble made sense :p
 
Back