P4 2.2GHz and D845BG Review

nihilist_1137 writes "GreenJifa.Com has gotten their hands on the new Intel P4 2.2GHz/Intel D845BG DDR Motherboard for review. This is the new P4 that has the 0.13m die and the new "Northwood" core. Check out the review." This setup might have a chance to run XP without it feeling like a 386/16 running Windows 3.0 on 4 megs of RAM. Allright, thats probably crazy talk ;)

What's the Hurry?!?

[The Gardener]

Four comments, none above zero, and its already Slashdotted

The Gardener

Well Thought Out?

[The Gardener]

This was the system I benchmarked the P4 on. I used 128MB of Micron PC1600 (200MHz) ECC DDR Memory.

The latest, preproduction, Intel CPU, and he only springs for 128 MB of ram? Why bottleneck the thing? No one is going to production ship it like that. I will likely go out the door with 512 or so.

The Gardener

Re:boot times

[Nameles]

Even in XP, I've seen the fastest bootup times of ANY OS (Win9x's, ME, 2k, Linux, Win 3.x, old old old OS's that I can't remember the names of) I've used.

Re:boot times

[trentfoley]

I use a laptop and am rather mobile with it. Boot times are very important when you just need to check that one little thing that some pesky client "needs". There is nothing more frustrating than having the conversation with the client go on to something entirely different while still waiting for a login prompt to appear.

Re:boot times

[ivarneli]

The operating system is not the only factor that decides system uptime. Many people do not leave their computer on all the time for one reason or another. For example, I turn mine off every night because I can't sleep with the noise. But boot times are important for many other reasons. Boot times are very important for people who dual boot and have to switch operating systems often. My work forces me to switch between Windows and Linux about 6 times per day. If Windows takes 5 minutes to boot, that's a half hour of time that has been lost. It's even more important when doing service work on computers. I used to work at a local tech shop, and the reboot times add up when you have to install a number of different applications or drivers. Also, we used to test parts, so if I had 50 video cards to test, that meant booting Windows 50 times. A great deal of time is wasted depending on how long it takes to load.

BeOS is probably the fastest-booting full-featured operating system, taking about 7 seconds to boot. DOS is of course much faster, although there isn't a whole lot for it to do. MS-DOS 5.0 with no autoexec.bat/config.sys presents a command prompt pretty much instantly after POST, even on a 486. I once got an extremely minimal Linux to boot in 3 seconds on a Pentium 90 when I was designing a car mp3 stereo. However, that was without any daemons running, no unnecessary drivers, etc... so it was not really a usable general-purpose operating system. In the Windows world, I would consider Win95 to be the fastest (contrary to every review and ad for Win98+ that claimed faster boot and shutdown times). A fresh install of Win95 on my Athlon 800 loads in under 10 seconds, although this starts going up when you install a real video driver, Internet Explorer 5, etc. Still, if you carefully monitor what gets loaded on startup, you can keep Win95's boot time under 30 seconds, which helps immensely when rebooting many times daily is necessary. Even in cases of normal use, a short boot time is a great convenience.

But maybe that's just me.

XP doesn't seem to be in demand here

[CatherineCornelius]

If you're going to bash [XP], bash it on the potential Digital Rights Management that was supposed to be introduced in XP, or on the product activation, or on any other Microsoft expansionist move. Bashing it for being slow is mostly just uninformed.

Fair point. I'd also like to bash it for being rather insecure, but I suppose I'll just have to stand in line behind all the other /.ers who want to bash it for being a Microsoft product.

But what interests me about XP is that so far there's no sign that the people at our office who use 98, NT or 2K want to upgrade. There seems to be a curious lack of keenness about this product. Perhaps it's the digital rights stuff. We developers run partly Microsoft, mostly Linux, so the people I'm talking about are the CEO, sales, marketing, legal, administrative and whatnot. They're more than happy with 9X or 2k, it seems. Or perhaps they're just scared to move in case they end up having to pay out for "upgrades" in the future.

I guess we'll get an XP machine eventually because we are an ISP and we will have to support our users, just as we run tests on Mac (9 and X). But somehow our people don't seem to see XP yet as a gotta-have, the way 95 was.

Re:Taco's XP comment

[peripatetic_bum]

Bottom line is that XP is no worse than any other "modern" graphical OS - it's just made by Microsoft. Accept the fact that Windows XP is a decent operating system and far superior to the Win9x line and get back to using your Linux PC.

I will have to respectfully disagree (unlike some of the replies ive seen to you)

We decided to set XP in my families new XP1800 computer. I wil be the first to admit that when it runs, it runs smoothly and the family likes it, but It certainly is much worse in terms of stability to say Mandrak or MacOSx. There is not a day that does not go by the computer iwll up and reboot for no reason or simply crash.

As for being a decent OS being made by Microsoft, all i can say is that we must remember that Microsoft KNEW about a HUGH REMOTE HOLE for almost a month before deciding to let the rest of us know about. That in my opinion makes it a very much worse OS than the others. At the moment, i am trying to get hem used to a linux desktop and will simply replace it all with Linux in a few months.

Thankx!

The end of interpreters!

[chrysalis]

Hardware is still getting faster and faster for the same price.

This is neat for developpers. Soon, source code will be recompiled in real time at every key stroke.

No more need for interpreters :)

Re:Taco's XP comment

[Pussy Is Money]

This is not true. Software may become faster as it evolves to exploit more powerful hardware. It may also become faster because of increased efficiency in the code.

But this does not mean that software gets faster with age. Because with age also come features and increased expectations. Imagine you had written a strcmp() routine ten years ago. You have had a decade to tweak your code and by now it is almost twice as fast as then.

However, now people expect Unicode support. If nothing else, this means your strcmp() has to compare twice as much data (not even counting character set translations and the like). So did your strcmp() actually become faster? Well, yes, it did -- but for any actual workload, no, it didn't.

I agree with your statement that we do not need a 2.2 GHz machine just to open a couple of applications. But people might require the ease of use (in the sense that the system has comprehensive knowledge about real world things such as character sets, physical dimensions, monetary values, etcetera) that a 2.2 GHz machine affords.

Re:boot times

[Drakino]

Then people need to look into sleep or hibernation. Everyone assumes it's for laptops, but it works just fine on desktops. I hibernate my media system all the time to have a middle of the road between power usage and boot times.

Macs should be using sleep mode if running OS X.

Re:Boot Time is Inversely Proportionate...

[Sycraft-fu]

Well a lot of that has to do with what is in your system. My PIII 700 with XP certianly boots slower than my 486 50 of yester year, however most of that is hardware. IT takes longer for the system to POST than for XP to load. IT has to scan 768MB ram, several HDs, 3 CD-ROMs, a SCSI card, etc, etc. It takes about 40 seconds to post and then around 20-25 for XP to load.

Now back when I had 2000, that took longet to load, it took like a minute or so, but now it's the POST not the OS that is the slower part. I tell you what, if you take a board with a celeron, 64 MB ram, a 4MB VESA card, SB16 and a single HD and boot to DOS, you're up and running in like 15 seconds flat :).

Re:As others will surely also state...

[Anonymous Coward]

Read the number of posts that use the exact wording, "Linux is not ready for the desktop." I am amazed that they are soooo obvious... Sad really.

And the tag team that two of them have going is also fun to watch... One makes a statement, then the other one supports it.

Tisk, tisk tisk. They must really be worried about XP getting trashed. We all know that sales of XP really sucked, and that this dragged down computer sales in the last quarter of the year. Maybe people wanted a choice as to what OS came on their machines.

Personally, I build my own machines. My favorite is a 2 year old dual celeron on a BP6, overclocked to 522MHz. It has 512 MB of RAM and 80 GB worth of RAID 5 hard drives. That machine is so fast, you click on a icon and the app pops up almost as you are letting up on the mouse. Even Mozilla starts in about 2 seconds.

I fix the boot sequence to boot to an Xwindows prompt in just a few seconds. All the servers and a lot of the services are started up after X windows is started. I am suprised that we don't have a better boot sequence in Linux yet. Especially since it was so easy to do this.

Maybe the distribution people need to get together and all agree to a new init method that emphasises boot speed for desktop users. Since this seems to be a problem for a lot of people here.

Only running applications that are linked against one set of libraries seems to help by not having to cache a lot of different libraries.

I also would like to see our applications get prelinked against the libraries that they are using, like under OSX. Linking at run time is just too expensive. It is better to do it just the one time and to save the executable prelinked. This can just be another step in the installation process. Prelinking would easily half the load speed of almost every program.

The last thing that I would recommend is using the intel compiler to compile a some of the executables that are taking a lot of processor, like audio/video codexes.

This would make them run much more efficiently and be able to take advantage of special instructions on the various platforms. This could easily result in upto 20% performance increase over gcc compiled executibles. Not because gcc is bad, but because intel is just good at writing x86 optimized compilers.

An increase of 20% will make a 500MHz processor run like a 600MHz processor, and this is the difference between dropping frames and not dropping frames during recording.

Can you imagine a Linux distribution that did these 3 simple steps? It would boot in 5 seconds to a login prompt, would start programs in sub second times and would need 20% less processor for the same performance as a normal Linux distribution. I'd pay good money for that distribution.