Slashdot Mirror
Intel Northwood CPU Review
gcshaw2nd writes: "Here it is, the first hands-on review I've seen of Intel's new Northwood chip, running at two gigahertz. It overclocks like a hog, easily to 2.5Ghz."
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
Aside from the meager "5-10%" performance boost per clock that GamePC reports, the new PC1066 RDRAM and 533MHz FSB coming in a few months offers a "12%" performance boost per clock, when used with the original P4.
Northwood + 533MHz FSB/PC1066 RDRAM should be quite nifty.
The PC1066 benchmarks are here.
According to that chart there, PC1066 RDRAM actually has lower latency than PC133 SDRAM. I don't know how accurate that is, but it says PC1066 RDRAM takes 207 cycles for 128 bytes, and PC133 takes 229 cycles (PC800 took 270)?
Maybe I'm reading that wrong or don't know some specifics about RDRAM architecture, but that sounds nifty...
Now I *know* my computer is obsolete: my CPU speed is just the difference between reccommended and possible clock speed settings on the new one.
Free Java games for your phone: Tontie, Sokoban
Some people have had that little gem up to 3Ghz. Not exactly in English, but pictures say a thousand words.
That's not true, they use RDRAM or SDRAM or DDR SDRAM. In fact there's a brand new artice that benchmarks them all.
Take your pick of RAM (although RDRAM and DDR are the only real choices).
When performing calculations that can take hours or days even, an increase in performance of even 10% can result in significant time/money savings. There are those mid level workstation users (like me) and high end users that can and do need every last bit of performance they can get. At this level, a few hundred $$'s every few months is nothing.
Hell, just the yearly support costs of a single SGI Octane are such that I could afford to purchase a new Macintosh G4 with a flat panel yearly for what it costs. So buying a new $800 chip twice a year does not even make me blink.
Visit Jonesblog and say hello.
now even quicker !! this page claims it has a world record
3023mhz !
John, Intel's VP of engineering : You mean "gigahertz" I 'spose : well, we can't, CPU core designs are reaching their limits. Technically, it's not very realistic you know.
Richard : Who's talking about technical stuffs here ? I mean, just crank up the gogohertz man, we need more hype fast.
John How about if we designed synchronous processors instead ? now *that* would be sexy.
Richard : You mean hotter than more gogohertz ?
John : Sure, it would impress the technical crowd, and we'd have a real actual breakthrough in CPU design in less than 5 year. That's pretty hot, I'd say.
Richard : Yeah, well, tech people are nice, but the Joe Sixpacks who walk into Fry's and buy a new PC, they want more gogohertz.
John : *sigh* Well, I guess we could double the clock and put a frequency divider inside the CPU
etc ...
When I bought my home workstation about three years ago, the CPU (K6-II/400) was one of the cheapest parts of it. What really made it powerful are the SCSI cache controller and a fast Seagate drive, as well as an adequate (for these days 128 MB were more than enough) amount of RAM. Of course it depends on what you do with your box, but to be honest, most of the time you're waiting for the harddisk, either for loading data or for swapping virtual memory.
Three years later, the only thing I added was some more RAM, with the rest of my workstation being the same. It is still very usable, and I rarely see the need for a more powerful CPU.
In contrast, my former office workstation was a P3-800 with 192 MB RAM (some of which to be "abused" for graphics), an IDE drive and a one-chip-does-everything Intel i810 on the mainboard. SETI was the only thing it could do faster. On pushing the IDE system or the network, sound playback got distorted, and the X server became quite unreactive, it even stalled for a few seconds. A compile run made it impossible to do anything different in parallel, so I would have needed two machines - one for compiling and one for the desktop.
AFAICT especially on Intel systems the trend goes towards integrated one-chip-does-everything systems like i810 and its successors, which can handle everything from graphics to sound to IDE to networking. Of course the Intel people want their customers to come back later, and save some money by using only one chip instead of several ones. Most users will think it's the CPU which is too slow... and buy a new 2 GHz monster with another i8xx-crippled mainboard.
Also To 2.8ghz
:)
But now, a little rant...
Look at the next and last alpha that will come out, it will SMOKE at a fraction of the speed the P4 will be at the end of the year. That's what I call going out with a bang. It's a shame such a technology got obliterated in part because of Intel's markettign muscle, we're going to be set back for 2 years for them to catch up on this technology and performance (they've got some alpha people in their staff now that will teach them how to do more than overclocking and recycling like putting SMT on chips for example, that will be one great leap, it's in at least one of the northwood flavor I think).
Watch out for the Hammer when it will come out, it still going to beat intel's latest offering mhz per mhz in desktop performance (can't speak too fast about the server side though.)
Look at the PowerPC, as much as I hate apple's hyping to keep their blinded userbase into beleiving that they hold the only computer that should exist on the surface of the planet, the powerPC architechture has proven itself to beat the crap out of Intel mhz to mhz side to side (and no I am not talking about that "hey loading an image in photoshop on the mac is 2x faster than on a PC, but I won't mention that it's totally unrelated to the CPU and the mac is running on a SCSI cheetah while the PC was running a 5400 rpm 5 gig drive", I am talking about rendering on different crossplatform software like premiere, lightwave, maya, etc).
This goes without mentionning SUN or MIPS or any other cpus on the planet that has interresting technologies ASIDE FROM CLOCKSPEED.
So again, Screw the MHZ hype, I am a power user, I love doing 3d rendering, I administer a small renderfarm at work, I love raw power, anything that comes out and had a power factor gets my attention, it's nice to see stuff running fast, but I am not impressed at ALL with the MHZ hype. Especially for the PRICE you have to pay to cover all the media and marketting hype... Intel's Hype tax like I often call it. Also, what you see is the MHZ going skyrocket... How much do you thihnk they had to cut in the design so that it stays that stable at these speeds? why do you think the athlon4 can't run at 2.2ghz? design.. intel had to cut in some places so the cpu could be easily cranked up that much, they had to redesign part of it, and that's why you need SSE2 optimisation and the pipeline for standard FPU is so bad. It's not because Intel doesn't know how to design a FPU, it's because they HAD to cut on it...
My precision 530 workstation that runs dual 1.7ghz P4 for the price I payed, I find a dual AthlonXP or MP 1900+ far more impressive for 1/2 of the price. Heck, you want to be in the cool factor? get a Dell 8100 with DVD/CD-RW Geforce2Go 32megs and 1600x1200 15" LCD screen, now THAT'S a nice little piece of technology. And quake plays soooo smooth on it... I won't waste my personnal money or blast my budget for a chip that can generate small blackholes because of it's so great CLOCKSPEED. Gimme raw speed at a decent price. For the price you'd pay that northwood alone you could probably get 2 athlonXP and it's mainboard that beat the crap out of the P4 for the same price of the cpu alone.
enuff
--- Metamoderating abusive downgraders since my 300th post.