Slashdot Mirror

Gigabyte is preparing an interesting solution. AnandTech give us a brief overview as seen in the last Computex:
http://www.anandtech.com/tradeshows/showdoc.aspx?i =2431&p=5

Basically they just use ordinary DDR and use a pack of batteries to keep the data when the computer is powered down. The batteries have a maximum life of 16 hours. So this is for enthusiasm that leave their computer always on. I wouldn't install an OS on this since in case of long power failre you would loose eveything, but I really wish I could have one so that I could install Battlefield 2 on this. It certainly would lower the very long load time. :)

http://news.designtechnica.com/article7516.html

It might help to read the original article. Note, for example, that the drive uses the "industry's highest density 8 GB" flash to create 16 GB drives, meaning the drive probably uses striping. Also, the drive's performance seems to be pretty good:

"The SSD's performance rate exceeds that of a comparably sized HDD by more than 150 percent. The storage disk reads data at 57 MegaBytes per second (MBps) and writes it at 32MBps."

Conservatively, that's right on par with the fastest non-SCSI drive in the world, and by the time it's released, it will probably be able to directly compete with 10,000 RPM SCSI drives. When you consider that this drive weighs half of what a regular hard drive does, uses 5% of the power, gives off minimal heat, and won't break if you don't treat it perfectly (I've had to bring my iBook into the shop twice, both because the hard drive broke), is there anything to complain about?

There's one important (MMORG) exception: Everquest 2. According to Anandtech, a GeForce 7800 GTX will only get you 25fps at 1280x1024 in extreme quality mode.

I wouldn't be surprised if Unfrozen Caveman Baseball Player already ordered a pair. He's on the disabled list and his 2005 salary is $14,500,000.

~70fps at 2048x1536 (four times 1024x768, 3x movie film speed).

Anandtech Doom 3 Benchmark

No, the RSX is G70-based.

From the Anandtech article:

"As we mentioned before, NVIDIA's RSX is the more PC-like of the two GPU solutions. Unlike ATI's offering, the RSX is based on a NVIDIA GPU, the upcoming G70 (the successor to the GeForce 6)."

The anandtech review addresses this issue.

It looks like ati is ahead on H.264 support.

http://www.anandtech.com/video/showdoc.aspx?i=2451 Anandtech has an excellent review which includes power consumption information and a good overview of technology in the new chip.

http://www.beyond3d.com/previews/nvidia/g70/Beyond 3D as always has a fantastic writeup including information on: CPU Utilization for video decoding, noise, power consumption, etc.

ExtremeTech has the first review

Anand's review was really good:

http://www.anandtech.com/video/showdoc.aspx?i=2451

HJ

Low end nVidia cards such as the 5200 you mention do a poor job with high-res 2d, though, which was the focus of the parent poster's question. In the 2d arena, ATI's offerings do a better job, apparently because there are fewer OEMs who use more consistently high-quality parts.

Goddamn you HTML.... http://anandtech.com/video/showdoc.aspx?i=2451

The opteron would be faster in about everything. Cheaper too.

clicky

AnandTech did this tour a few years ago I think.

http://www.anandtech.com/mb/showdoc.aspx?i=1891

HJ

Well, why don't you provide links to reviews?
Because they don't back up what your saying. I searched Anandtech, and the only recent review which showed up pitted a 2.0 Dothan against a 2.2 AMD64. And the Dothen lost most benchmarkds.

http://www.anandtech.com/mobile/showdoc.aspx?i=222 5&p=15

The well know one where the Dothan won a lot of benchmarks pitted a massivly overclocked Dothan against a FX and PEE running at stock speeds, at Tom's Hardware Guide. Hardly a fair comparison. If the FX, not to mention the PEE had been overclocked to the max, they would have wiped the Dothan. Not only that, the FX was using an old core. Naturally Tom's Whoreware guide did everything it could to slant the results in Intels favor, something that site is notorious far, witness the recent stress test.

But to paraphrase Harry Truman, their are three kinds of lies: lies, damn lies, and benchmarks. Over the last couple of years, as the A64 has become dominant, many benchmarks the P4 used to win in but the A64 took the lead in have disappeared. Just about every site is bought off by Intel to some extent, either conciously or unconciously by the huge marketing and advertising dollars Intel throws around. Its easy enought to find programs optomized for Intel architecure and synthetic benchmarks slanted to favor Intel. Unfortunately for Intel, AMD has taken such a lead, that doesn't even work anymore. And review sites, in order not to loose total credability can only distort so much without destroying their reputation, as Tom's Hardware as already done.

As for the Dothan being faster clock for clock, well duh, that doesn't mean squat. Much as the Athlon Barton XP being faster then the P4 Northwood clock for clock doesn't mean squat. If an architequres is designed to clock higher, and will do so, its faster, period.

That being said, I agree the Dothan is a very good chip, and is competive with the A64 in gaming benchmarks, but not in others, such as media creation, much the same as the P4 is dominates in some types of software over the A64. Intel would do well to bring it to the desktop and tweak it to its full potential.

Bullshit, Intels' Dothan (Pentium-M) on an Asus mobo will smoke an FX-57 at less that half the price. Dothans currently hold all the 3DMark records and SuperPi

The article said that the FX-57 crushes everything in its path in game performance. You bring up the Dothan. The Dothan will not "smoke" a FX-57. While it is a fast chip, it is not a desktop replacement chip and lacks the power that the high-end Athlons have in games.

http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2382&p=7

I'm not enough of a Big Iron guy to know if there's rampant holes in these benchmarks or not, but this benchmark set at Anandtech (and other pages in that article) suggests the Mach kernel in OS X isn't the greatest for high end server stuff. So is this the smartest move Apple could make?

As for the viability of WebObjects, well... I'm currently a J2EE developer working with in-house libraries. Once I get my thesis written I'm going to spend some time with one of the next generation web development platforms; either some more Java libraries (Spring/Webwork/Hibernate), Rails, or Seaside. I'm afraid WebObjects is a good long way down my list, and I'm a daily OS X user! I'd maybe have thought about it if I could have rolled out onto Tomcat, but now I can't, well, it appeals to me even less.

A review of these chips compaing performance on AMD and Intel processors!

Perhaps you should read this before moving to OS X.

Performance is at that point only 1/10th of the Opteron and Xeon. We have tested this on Panther (10.3) and on Tiger (10.4.1), triple-checked every possible error and the result remains the same: something is terribly wrong with the MySQL server performance.

SPEC CPU 2000 Int numbers compiled with GCC show that the G5 reaches about 75% of the integer performance of an equally clocked Opteron. So, the purely integer performance is not the issue. The Opteron should be quite faster, but not 10 times faster.

We checked with the activity monitor, and the CPUs were indeed working hard: up to 185% CPU load on the MySQL process. Notice that the MySQL process consists of no less than 60 threads.

But, as I talk a bit more about elswhere, for most people it doesn't make half as much sense as using a Darwin distro like OpenDarwin on your servers for OS X client networks.

I'm not quite so sure I agree. As the recent Anandtech article showed, the kernel in OS X and Darwin can really bog things like databases and web servers under excessive connections, due to the coarse threading inside Mach.

Besides which, there is a much larger choice of server applications on Linux. Need to run DB2? Not goinng to happen on OpenDarwin. Or Novell's cool administrative tools? Linux again.

That's not to say that I have anything against Darwin and its derivitives -- I'm technically running Darwin as part of Tiger. I just think you'd have a lot of work to do to convince me that OpenDarwin makes a better small server choice for OS X networks than Linux.

Yaz.

That is, unless Apple fixes this little issue outlined in this article:

http://www.anandtech.com/mac/showdoc.aspx?i=2436&p =7

(disclaimer: i'm an apple fan and was disappointed to see this)

From Anandtech.com ( http://www.anandtech.com/video/showdoc.aspx?i=2406 )

ATI X850 XT

Doom 3:
Frame rate 1024X768 = 97.6
Frame rate 1600X1200 = 52.9
Frame rate 1024X768 (4XAA, 8XAF) = 65.0
Frame rate 1600X1200(4XAA, 8XAF) = 34.3

Splinter Cell Caos Theory:
Frame rate 1024X768 = 79.9
Frame rate 1600X1200 = 46.0
Frame rate 1024X768 (4XAA, 8XAF) = 66.1
Frame rate 1600X1200(4XAA, 8XAF) = 36.6

Half Life 2:
Frame rate 1024X768 = 145.6
Frame rate 1600X1200 = 116.9
Frame rate 1024X768 (4XAA, 8XAF) = 140.5
Frame rate 1600X1200(4XAA, 8XAF) = 86.0

What you'll notice across all tests is that you can run Anti-Aliasing and Antriscopic Filtering at their highest levels at a lower resolution and still maintain a higher framerate than by simply increasing the resolution. One could assume (although it may not be true) that if you were to run your game at 800X600 with 4XAA and 8XAF that the game would maintain a higher frame rate than the 1024X768 resolution without AA or AF.

Of course price isn't everything to them, but it is one thing as Intel has a very mature manufacturing process. Apple especially found the low power, high performance Pentium M a very attractive option. In fact, one may argue that it's the 'final push' that brought Apple to Intel. Also Apple was having some difficulty with their PowerPC venders:

- IBM couldn't get the power consumption of the G5 down, which is why we haven't seen a G5 laptop - IBM was having yield problems with their G5 model - The G5 wasn't up to snuff with it's x86 competitors source here - Freescale Semiconductor, Apple's other business partner who designed the G4, is primarily interested in the embedded market

This is what I know. ^_~

The line in the 'about' box said 'Pentium 4-processor', and the '-' was in the wrong place - it should have said 'Pentium-4 processor', but it *was* a pre-production machine for developers only... this audience is supposed to have a clue...

Looks to me like there's no hyphen at all.

1. What happened to the PowerPC's supposed performance advantage over Intel?
   Absolutely nothing. The article , and it is unlikely that Apple will suffer greatly from this. They have other products such as their iPod and iTunes services to support themselves. Sure, sales will fall, but it's my prediction that AAPL will fall and then pick up as market analysts predict a rise in Apple sales in the next few months due to a new product release (Intel Macs). The Osborne Effect doesn't really hold water, Apple already have a development system available, and have already ported their OS. They have been planning this for five years. They do have a product to deliver, and they are very, very good at hype.
2. Is this all really about Digital Rights Management?
   He's right on this one. No.

1. What happened to the PowerPC's supposed performance advantage over Intel?
   Absolutely nothing. The article he refers to in Question 2 answers his question here. The introduction of the PPC 970MP with a 90/65nm fab process would allow the G5 CPU to hit 3.5 GHz and use less power too. This wasn't bullshit. The G5 was clearly faster for raw calculating power (agreed, the linked article shows some dire results for MySQL and so on, but this is more likely down to how the OS handles threading, or how MySQL was compiled).
2. What happened to Apple's 64-bit operating system?
   Nothing. I assume that the new Apples will not use Xeons or Itaniums, but Intel's next desktop chip (Pentium D?) with AMD64/EM64T 64-bit extensions.
3. Where the heck is AMD?
   AMD's fab plants are running to maximum capacity, as are IBM's (all next gen consoles are using IBM's chips). They are not the sensible choice. Intel has the capacity and the know-how. Apple are also free to switch to AMD if Intel turns out to suck, although this will cause another uproar.
4. Why announce this chip swap a year before it will even begin for customers?
   To prepare corporate customers and their user base for the switch. To give developers time to port software to the new architecture so that it will be ready on release of the new system. Cringely's answer to this question is stolen from The Register, and it is unlikely that Apple will suffer greatly from this. They have other products such as their iPod and iTunes services to support themselves. Sure, sales will fall, but it's my prediction that AAPL will fall and then pick up as market analysts predict a rise in Apple sales in the next few months due to a new product release (Intel Macs). The Osborne Effect doesn't really hold water, Apple already have a development system available, and have already ported their OS. They have been planning this for five years. They do have a product to deliver, and they are very, very good at hype.
5. Is this all really about Digital Rights Management?
   He's right on this one. No.

One example of huge gains for a lot money:
http://www.anandtech.com/video/showdoc.aspx?i=2278 &p=3

Compare the NVIDIA GeForce 6800 GT 256MB, being at $295, and a comparable NVIDIA GeForce 6800 Ultra 256MB, at $419. For an incredible gain of 8.5 frames per second! That's %7 percent performance increase for only a %42 price increase.

Hey... if you have the money to waste, though... be my guest and waste it.

This recent AnandTech article was very disappointing to read as an Apple (and *nix) fan. Basically, the more requests were thrown at one time at an Apache or MySQL server, the faster Linux looked compared to OS X. In fact, OS X pretty much got clobbered with an increasing number of requests... apparently because of the threading model underneath and/or the wrappers around it. I would hope that Apple finds a way to alleviate that situation ASAP. In fact, I can't believe they're even able to sell OS X Servers if this situation is as real as this article makes it seem, unless there's something I'm not aware of.

"I like the fact that both of my macs running together at full noise are quieter than my PC with one fan. This is because of the CPU. PowerPCs are cooler in both senses of the word."

Maybe you should check with Steve before you continue spouting propaganda. Apple wants cool and quiet computers, so they picked Intel.

Pentium 4s are rediculous, but as far as I can tell Apple will be using Pentium M derived chips for everything except those developer boxes they're shipping now. Pentium M blows Apple away in terms of battery life, they are easily the cooler chip.

Already done, very recently

http://www.anandtech.com/mac/showdoc.aspx?i=2436

Except it's dual AMD64s @ 2.4Ghz vs. G5s at 2.5Ghz and 2.7Ghz

The short story: The G5s get stomped all over the place, especially in server tasks.

Please read this. and this Thanks

http://anandtech.com/tradeshows/showdoc.aspx?i=243 9

You can buy them now: http://www.anandtech.com/news/shownews.aspx?i=2437 3 HJ

Well, for one, they don't work together as well as the AMD. With the Intel, any communiction between the cores has to be done through the FSB, where as with AMD, its done internal to the processor. That is huge. The on-die memory controller on the AMD just makes the chip much more advanced than anything Intel has.

One big advantage AMD has is that as you add cores, the performance hit you take over contention for memory is very low. For example, look at these benchmarks: Opteron vs Xeon

The Operton in a 4-way configuration makes a mockery of the Inetl Xeon and its running 800Mhz slower and with much less power consumption. People updating from dual opteron to quad by going multicore are going to get a lot more bang for their buck, specially since they don't need to buy a new motherboard.

Well, for one, they don't work together as well as the AMD. With the Intel, any communiction between the cores has to be done through the FSB, where as with AMD, its done internal to the processor. That is huge. The on-die memory controller on the AMD just makes the chip much more advanced than anything Intel has.

One big advantage AMD has is that as you add cores, the performance hit you take over contention for memory is very low. For example, look at these benchmarks: Opteron vs Xeon

The Operton in a 4-way configuration makes a mockery of the Inetl Xeon and its running 800Mhz slower and with much less power consumption. People updating from dual opteron to quad by going multicore are going to get a lot more bang for their buck, specially since they don't need to buy a new motherboard.

An Anandtech review concludes that the G5 is somewhat crippled by Mac OS X when used for multi-threaded server applications like MySQL. The reviewer recommends serving with Yellow Dog Linux over Mac OS X for this reason.

Presumably, the Intel version of OS X has the same threading and locking problems inherent to Darwin. Until they fix this, there'll always be a place on an Intel Mac for Linux development.

Until Tiger came out I would agree that statement. Tiger changed long standix POSIX networking API's. I upgraded to Tiger and nothing but the very trivial of apps runs. Gnome to took forcibly removing gnome-vfs-ssl and replacing it with gnome-vfs. Gnome apps still take about 1 - 2 minutes to display on the screen after which they run fine except for gnome-terminal which doesn't work at all. There are no kde apps in fink for Tiger due to message passing being crippled in kdeinit. The apps compile but nothing except nedit, grace and gimp seem to run normally at all and gimp has slowed considerably. I've been waiting for the fixes in fink but they have been extremely hampered.

And according these benchmarks Nobody would consider OS X a superior server solution.

OS X in Intel will only emulate OS X in G3. Not G4 or G5, not Altivec, not OS 9 and not 68K.

Anand coveres it pretty well here.

Good to see you've got your facts straight before you posted.

AMD does not have a CPU running anywhere NEAR 4GHz, you're thinking of Intel.

As far as power consumption..
"Even the Athlon 64 X2 4800+ consumes less power than all single core 90nm Pentium 4 CPUs" - Anandtech

For more information please see this and this

For less power, better performance use AMD.

Good to see you've got your facts straight before you posted.

AMD does not have a CPU running anywhere NEAR 4GHz, you're thinking of Intel.

As far as power consumption..
"Even the Athlon 64 X2 4800+ consumes less power than all single core 90nm Pentium 4 CPUs" - Anandtech

For more information please see this and this

For less power, better performance use AMD.

someone did that comparison, linked here

So not only has Apple dumped IBM, they also appear to be planning to dump gcc.

Maybe my link pointed you too far into the article. Try starting at the beginning instead. Sorry about that. Anand does compare workstation performance which in many cases is pretty close to the competition. Only rarely does the G5 seem to win. I looked pretty closely at those Lightwave comparisons, since I actually use that app.

While over there don't miss his article on this very issue. Nice coverage as always from Anand.

Maybe my link pointed you too far into the article. Try starting at the beginning instead. Sorry about that. Anand does compare workstation performance which in many cases is pretty close to the competition. Only rarely does the G5 seem to win. I looked pretty closely at those Lightwave comparisons, since I actually use that app.

While over there don't miss his article on this very issue. Nice coverage as always from Anand.

It would be interesting to see some benchmark comparisons though. I mean how sure can you be of something that just feels faster? Try a few cross-platform benchmarks. Although you also have to consider which Pentium M you have. As far as heat goes, that's really hard to judge because it depends so much on your respective cooling systems. From what I've read about the average TDP for both chips, the Pentium M wins on paper, but I've never seen any published benchmarks between a Pentium M and a G4. And those TDP numbers are not maximums. So there's lots of wiggle room for the marketdroids on all sides.

Anand's recent benchmarks seem to disagree with you. As well as Steve himself, who characterizes Intel as having better performance per watt. And indeed nothing can touch the Pentium M in that respect.

One thing you have to keep in mind is that the current Apple user base is minuscule compared to the Wintel crowd.

I'm sure Jobs knows this and is trying to build Apple and put them in a power position. It's not that they don't care about their current user base, but the fact of the matter is that they just don't matter compared to the massive number of people they will gain by going x86.

As a side note, I'm not entirely happy with how OS X is designed. It's the whole problem of its lackluster performance (Aqua on top of Darwin/BSD on top of Mach gives you a nasty penalty; just look at the benchmarks).

You value battery life, but you bought a "mobile" Pentium 4 laptop instead of a Pentium M or Celeron M laptop at the end of 2004? It's not like the power-saving features of Centrino haven't been publicized and hyped since March 2003.

Initially, Pentium M/Celeron M notebooks carried a notable price difference over Pentium 4/Celeron notebooks. But by the end of 2004, Pentium M technology had made its way all the way down to sub-$1000 notebooks (Celeron M). By that time, only "desktop replacements" (and older discontinued models) were using Pentium 4 CPUs. For the same price as that Toshiba Pentium 4-based notebook, I think you would have been much better off (with much longer battery life) with a Pentium M-based, or even Celeron M-based, notebook.

Working link...

It's not official until Steve Jobs casts his distortion field over the WWDC attendees on Monday and tries to explain to them why going with the Itanic is a good idea.

AMD would have been a MUCH better partner, IMHO, as Apple will be "just another partner" to Intel, and the center of attention at AMD.

Anyways, Apple needs to stop complaining about chip problems and fix their P.O.S. kernel. Did you see the benchmarks from this article?

Good Lord does OS X suck as a server.