I’m planning on buying a laptop, and I have a question about Intel Centrino. What exactly does it do? Does it provide wireless Internet service? Does it cost a monthly fee to have the wireless Internet?
Basically, Centrino can just be viewed as a collection of “standards”, that Intel has put together in order to maximize the performance of mobile computers.
One of the major things you will notice about Centrino laptops is the lower clock speed of the CPUs. You can find some laptops with clock speeds in the 2.0GHz range and up, but a (comparable) Centrino certified laptop might actually run around 1.7GHz or so. More clock cycles = more power consumed. So if we reduce the clock cycle, we can extend the life of laptop batteries (before recharging, that is).
Some people might look at this as a hindrance of the laptops; lower CPU speed = lower performance, right? Nope. Take a look at the Athlon processors put out by AMD. These processors are able to compete, performance-wise, with Intel processors, even though their clock speeds are considerably lower (when Intel was releasing their 2GHz CPUs, AMD was around 1.4GHz, IIRC). How can this be? Basically, because the AMD processors will handle more “instructions per [clock] cycle” than their Intel counterparts.
And that is pretty much what happened with the Intel Centrino CPUs. The architecture was revamped to accomodate more “instructions per second” than the typical Intel CPU.
Of course, the other major thing about Centrino certified laptops is: WiFi capability. Centrino laptops must have a wireless adapter of some sort (802.11a,b,g…etc.). This (obviously) helps create a more mobile computer (no attached wires).
In order to utilize wireless internet, two main things need to be present: a client (a Centrino laptop, for example… or anything with a wireless adapter) and a server (aWireless Access Point - WAP - that broadcasts the network’s radio signal). So, as far as providing “wireless internet access”… it’s not quite as straightforward as that.
There are some services out there (T-Mobile, for example) that will allow you to use their wireless access points, for a monthly subscription fee (or pay-by-the-minute setups, in some cases).
I hope that helps.
LilShieste
A Centrino laptop is pretty much just a laptop that contains a Pentium-M chip (not to be confused with a Pentium 4-M; Pentium Ms use less power, and are clock-for-clock much faster than a Pentium 4.) and has a built-in 802.11 type wireless adapter.
The wireless adapter lets you hook into a wireless network. For example, my parents recently set up a wireless network in their house so they can use their Centrino laptop & desktop machines on the net at the same time. Setting up a wireless network at home is pretty easy(my parents did it, and they are not the best with computers); they just connected a wireless router (available for $40-$60 at your local electronics place) to their cable modem with an ethernet cord, and then connected their desktop machine (which sits next to the router) into the router with another ethernet cord. (Most wireless routers have a couple wired ethernet ports to hook up machines that lack a wireless adapter.)
Then they ran the wireless setup on the router, then ran it on the laptop, and wham! they could browse the net from anywhere in the house on the laptop. They later added in another desktop computer; they had to put a wireless adapter card in it, but after that, they could have all three machines on the net at the same time.
If you are on the go, you can get internet access from a laptop with a wireless adapter (like the Centrino ones) by either paying for the service in some areas (like that T-mobileLilShieste mentioned) or by finding a public wireless access point. Some businesses(like some Starbucks) will set up free wireless access points to draw in customers. Some people will open up a wireless access point out of the kindness of their heart. Other people will forget to set up security on their personel wireless routers, and anyone will be able to access it to use their network. I had a neighbor who did this once; I did the ethical thing and tracked him down and told him about it, and then helped him set up the security measures.
Everything I read about the Pentium-M suggests that its pretty much a Pentium III core, with the ability to use DDR memory.
Even more interesting is that clock for clock, the PIII was always faster than the PIV and used less power, and that the Intel version of multicore processing will use the Pentium-M as the underlying architectire.
Another advantage is the built in wireless antenna. My wireless card (in fact pretty much any add on wireless antenna) sticks out of the side of the laptop or involves a USB slot. Having the antenna built in makes the machine sleeker and reduces the possibility of damage.
Centrino is essentially the brand name for Intel’s 802.11b/g (also known as wireless, WIFI, etc.) technology. That’s it. Other than the Intel name, it gives you absolutely zero advantages over a wireless chipset made by another company.
Intel only allows companies (like Dell, HP, etc.) to market laptops with the Centrino name if they have both a Pentium-M processor and a Centrino chipset, but that’s just a marketing gimmick. You can buy Pentium-M-equipped laptops regardless of who manufactured the wireless chip they use.
This is not to say that you should avoid buying a “Centrino” laptop, as there’s nothing wrong with them, but they don’t really have any advantages over the (often cheaper) laptops without Intel’s hype-generator in them.
You don’t need to buy a Centrino laptop to get a built-in wireless antenna. Any laptop equipped with wireless from the factory* has that, Centrino or not.
The Pentium M shares many important architectural similarities with the P3 but was designed from the ground up as a purely mobile processor. The fact that P3’s could only ever reach 1.2Ghz or so stock while P-M’s are moving up to well over 2.0Ghz without showing signs of stopping shows that there were significant architectural differences between the two.
…erm except of course that I’m running a PIIIS 1.4Ghz and since it has a larger cache than the PIII 1.2Ghz it is significantly faster - not just incrementally, and it runs just fine at 1.5Ghz(on bog standard cooling - as it has for a couple of years now), it does mean I have to use PC 150 memory, just cannot get hold of the PC 166.
I’ve seen articles claiming the same core was capable of running at 1.8Ghz or more but that intel didnt want it competing with the early P4s, as it was already faster clock for clock, and runs cooler doing it.
Intel wanted us all to upgrade m/board, memory and cpu all at once, just like they’d forced users to do before(remember slot 1 then MMX & slot 2 and then FCPGA all in around a year or so? remember the joke that was RIMRAM?), and its why AMD is now emerging as a real competitor as AMDs upgrade lines tend to be both longer, and cheaper in that you usually have less to change fewer components step up one level.
One reason the PIII actually did more work per clock cycle is that P4 has that long long pipeline, which should make things faster if every instruction that was prefetched was actually required, however when its not, the pipeline ends up stalling whilst the right instruction is retrieved, they still haven’t sorted out the prediction properly so that the pipeline gets stalled all too often. The PIII has the same problem except that the pipeline is so much shorter and doesnt get bogged down for as long and the number of instructions that are prefetched are fewer, therefore the retrieval of the wrong one is less likely.
It wouldn’t be very good for intel PR to admit that the PIII was a better route to take than the P4, had it been redesigned to use DDR memory instead of going for P4 and had they also allowed it to reach its potential clock speed, maybe AMD would not have had the chance to jump in with cheaper and faster systems.
In effect Intel is admitting this by using major parts of the PIII for both the Pentium M and also for multicore processing where heat dissipation is a problem.
The Pentium M is in many ways the cpu that intel should have produced instead of the P4, but they got their strategy wrong, and took their customers for granted, expecting them to shell out yet again for virtually the entire heart of the system, and they haven’t stopped doing it either, look at socket 775, if you want it you need m/board, expensive memory, new graphics card since its PCIE only and not AGP, and very likely you’ll need to replace your hard drives too as there seems to be no support for IDE, SATA only.
Yet with AMD, all I need do is replace the m/board and cpu (if I already have an XP CPU) and I can have 64 bit system which will knock the spots of the XP CPUs, and I can do it for a fraction of the cost of trying to upgrade from P4 to socket 775, what’s worse is that soon Intel will be bringing out a new m/b format that will also mean you need to replace your power supply and case too.
Intel just do not learn.
Up to now I have been very happy with my Intel based systems but, I won’t be going the Intel path again, not until they take a look at the whole life cost of computing.
Faulty recollections here,
The MMX fiasco was during the Pentium I years IIRC Intel brought out the PI 166 and within a couple of weeks then followed with the MMX so that those who’d paid a premium for the 166 were left holding a less desirable chip in very short order.
IIRC it was around this time or just before that the PII was released, on a differant motherboard and without the MMX instruction set.
The process used to make the PIII on the Tualatin core was certainly good for 1.8Ghz and had it been taken on to the sub .13u fabrication would have gone on to higher clock speed, which is one of the main things they’ve done with the Pentium M.
Take a look at some cpu history.
There is a difference between the P3 and P4 architechture but it’s impossible to say which one is “better”. Each has it’s own strengths. The P4 architecture with the longer pipelines allows it to reach much, much higher clock speeds, 3.6Ghz as compared to 1.4Ghz which means that it’s undeniably faster regardless of what IPC can do. Also, the workflow for the different chips is different. Intel strongly banked on the fact that media production would be major force of demand for high end CPU’s in the next 5 - 10 years which requires huge amounts of bandwidth and CPU cycles. The pipeline length doesn’t matter all that much since branching happens so rarely and it shows whenever you see benchmarks of the P4 in media encoding situations. As a desktop architecture, it seems to be moving in the right direction. Using a P-M in the desktop would have been a much, much less successful move. The P-M is expensive to manufacture due to the huge amounts of L2 cache and it just simply cannot ramp up as high as the P4.