Print 10 comment(s) - last by Anonymous Frea.. on Apr 9 at 2:26 PM

In a bid to make manufacturing of Celeron M 500 CPUs economically more efficient, Intel will introduce a shrunk-down core in the near future.

Last January Intel introduced the Celeron M 520, a low-cost entry-level mobile CPU based on Core microarchitecture. At $134 the CPU was relatively low-cost from a consumer point of view -- with scheduled price drop to $107 on April 22nd. Under the hood the CPU featured a "fully fledged" Merom core with the current B2 stepping. It's safe to assume that these feature at least 2 cores and 2MB L2 Cache -- given the processors have different manufacturing lines for 4MB and 2MB models.

In the recent PCN 107423-00 Intel notified its customers about a stepping change for Celeron M 520. Although a Celeron M 530 was introduced on March 25th at $134, in the PCN this model is not mentioned at all. Intel informs that said model will transition from a Merom B2 stepping core to a new Merom-L core with A stepping, which will be available starting the 1st of June.

The only change for customers is a different CPUID which requires a BIOS update along with a microcode update in oder to support the CPU correctly. Besides Intel notes that "Merom-L A stepping material has minor die size decrease due to manufacturing optimizations to
increase product availability".

As the PCN only mentions a minor die size decrease it remains open whether Intel only reduced the L2 cache down to 1MB or also managed to cut off one of the two CPU cores. My bet is on L2 Cache only, as it explicitly mentions "minor" and the CPU cores in Core-based CPUs are linked kinda tightly so it would have required some sort of redesign of the chip.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

By StevoLincolnite on 4/8/2007 12:36:45 AM , Rating: 4
I'm actually starting to like the celeron now, I never liked the Celeron A series, Except for the 300A because it could do uber overclocks, But when they released the Celeron based off the Pentium 3 die, it wasn't all that great. I think the Celeron started to shine when they brought out the Celeron M based off the Dothan core which had 1Mb of cache, And it actually didn't perform that much different from my Pentium M.
I think Intel learnt that having a processor on a restricted FSB for so long is a bad idea, And when I look for my next system I'll probably look towards the celeron as a cost alternative, Besides I could save the money I save from buying a celeron and putting it towards and new graphics card :) Then later on upgrade to Core 2.

RE: Celery!
By tehfire on 4/8/2007 1:40:16 AM , Rating: 2
fdqwI don't think it was the restricted FSB so much as it was the lacking L2. An inclusive cache just does not take an L2 cache decrease very well. Of course, then decreasing the FSB bandwidth just when the processor needs it most doesn't help it either...

Typing with 4 fingers, football accident. Really lazy in typing and not explaining. Why do we geeks have to be so clumsy? :-P

RE: Celery!
By Ringold on 4/8/2007 4:16:08 PM , Rating: 2
Geeks play football..?

Just teasing :)

RE: Celery!
By Anonymous Freak on 4/9/2007 2:26:36 PM , Rating: 2
It's almost certainly the cache. I had a 1.4 GHz Celeron-M notebook (1 MB L2 cache,) that runs circles around a 2.8 GHz Celeron-D desktop (256 KB L2 cache,) with otherwise identical specs. (Both 915 chipset, both 512 MB dual-channel DDR2-533 memory.) And the desktop has an 533 MHz bus compared to the mobile's 400 MHz bus. I know the Pentium-M core is more efficient than the NetBurst core, but to the tune of more than double? The larger cache HAS to be the difference. I later upgraded the notebook to a Pentium-M 2.26 GHz and there really was very little noticable performance improvement, it took benchmarks to see the difference. When I upgraded the desktop to a Pentium 4 3.2 GHz, on the other hand, it was night-and-day.

RE: Celery!
By cocoviper on 4/8/2007 3:05:28 AM , Rating: 2
The 667 celeron was great too as it would easily ratchet up to a quick 1GHz when those CPUs were pulling $600 a pop and only for about a $100 chip!

As for the celeron's "bad" performance, the celeron was actually a very good performer with the exception of two versions: The Pentium II celeron (pre-A versions ala 300 vs 300A) that had zero L2 cache (just stupid with the P6 core) and the initial P4 celerons (which had their cache associativity cut way down).

RE: Celery!
By StevoLincolnite on 4/8/2007 7:17:29 AM , Rating: 4
Actually doing some tests I found out that a Katmai Pentium 3 500-512k of cache-133mhz FSB performed about the same as a 700Mhz Celeron.
The celerons not only were restricted to a 66Mhz bus in those days, Where the Pentium 3 used 100/133mhz buses, (Thats like 50%-100% bandwidth difference)
But they used 128k of on-die cache, Which was half of the coppermine and 1/4th the amount on the katmai, Now not only did it have its cache severely cut done, but the celeron only had four-way set associative on die L2 , While the Pentium 3 had 8-way, And cache latency's were much higher.
In 3D mark 2000 the Pentium 3 at the same clock speed as a celeron with a 100mhz FSB performed about 1000 3D marks slower. Which was really noticeable for back then. And back then, I prefer the AMD K6-2 Socket 7, then moved onto the K6-3 then the Pentium 3.
Back when the Celeron was against the Pentium 2, The cache was actually in a better configuration, As it was full speed on-die, Whilst the Pentium 2 was 256/512 off die cache.

RE: Celery!
By Lonyo on 4/8/2007 4:52:13 AM , Rating: 2
Well, I'm not sure if it's still true, but IIRC some of the earlyier Celeron-M's didn't have power saving features (Speedstep) which means they are not hugely beneficial for laptops, assuming you want a decent battery life. I could be wrong though.

RE: Celery!
By StevoLincolnite on 4/8/2007 5:38:16 AM , Rating: 2
No they didn't, yet they still consumed less power than a Pentium 4-M, I used to get 3 hours out of my Acer Extensa 2300 (Celeron M 1.5ghz) And there are programs around where you can under volt them.

RE: Celery!
By ChronoReverse on 4/8/2007 5:18:41 PM , Rating: 2
The Tualatin based Celerons were also good. They had less cache than the Tualatin PIII's but the same amount as the Coppermine ones. The latency was 1 instead of 0 but really who would notice that on a day-to-day basis? Plus the 1GHz easily overclocked to at least 1.33GHz =)

New launches
By crystal clear on 4/8/2007 12:19:48 PM , Rating: 1
Core 2 Extreme QX6800


Intel quietly launches first Core 2 Duo ULV processors

Santa Clara (CA) – Intel has added two ultra low voltage versions of its Core 2 Duo processor with Merom core to its product portfolio.

The new processors, the 1.06 GHz U7500 and the 1.2 GHz U7600, are based on the Core 2 Duo processor with 533 MHZ FSB and 2 MB L2 cache. However, the dual-core is rated at a thermal design power of 10 watts and will be targeting mini notebooks, tabl;et PCs and ultra-portable notebooks.

The U7500 and U7600 are introduced as part of the currently shipping “Napa refresh” platform, but will also be available for the upcoming Santa Rosa platform with a slight change in pinout (Santa Rosa processors use a P pinout instead of the M pinout of the Napa refresh processors.) The new Merom ULV CPUs will be replacing the Core Duo U2500 processor over the coming months.

One of the first manufacturers offering the U7500/U7600 will be Gateway who said that its E-100M notebook will be available with the two processors beginning on May 3 of this year.

"My sex life is pretty good" -- Steve Jobs' random musings during the 2010 D8 conference
Related Articles

Most Popular ArticlesAre you ready for this ? HyperDrive Aircraft
September 24, 2016, 9:29 AM
Leaked – Samsung S8 is a Dream and a Dream 2
September 25, 2016, 8:00 AM
Inspiron Laptops & 2-in-1 PCs
September 25, 2016, 9:00 AM
Snapchat’s New Sunglasses are a Spectacle – No Pun Intended
September 24, 2016, 9:02 AM
Walmart may get "Robot Shopping Carts?"
September 17, 2016, 6:01 AM

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki