Dell Dimension XPS: Gaming with all the Trimmings
by Wesley Fink on November 13, 2003 10:35 PM EST- Posted in
- Systems
Dell Dimension XPS: The Test
The full suite of benchmarks was run with the Dell Dimension XPS gaming system. Dell shipped the XPS with an ATI Radeon 9800 XT video card, which was used for all benchmarks with the Dell system. Please keep in mind that all of our comparison benchmarks were run with the ATI Radeon 9800 PRO, which is slightly slower. The 9800 PRO and XT use essentially the same GPU, but the XT is clocked a bit faster than the 9800 PRO. For a better idea of how the two cards compare, please refer to our benchmarks in AnandTech's ATI 9800 XT review.Memory timings are not adjustable on the Dimension XPS, so benchmarks on the Dell were run with the memory and setup as provided by Dell. We confirmed with CPU-Z that Dell runs the installed DDR400 memory at 3-3-3-8 timings. This compares to the 2-2-2-6 timings that we normally use with Mushkin PC3500 Level 2 or OCZ 3500 Platinum Ltd in our system reviews at DDR400.
Dell provided a complete installation on the test XPS and we made no attempt to reinstall. To provide the best test results under these circumstances, all co-resident applications and special services loading at startup were turned off so they would not load at boot. Benchmarks were installed and run from the nearly empty 500GB SATA RAID array. AnandTech benchmarks are normally run with sound disabled to remove the influence of varying overhead depending on the sound chip used on the board or video card. We, therefore, disabled all on-board sound on the Dell Dimension XPS to provide the most comparable benchmarks.
Dell provided the test XPS with an installation of XP Home, rather than the XP Professional, which is our normal standard. Since Dell considers the XPS to be a high-end machine targeted at the serious gamer, we compared the performance of the XPS to other machines and boards that we have tested and might be considered by a serious gamer. We included results from the Asus P4C800-E, a top 875P motherboard, and Athlon64 FX, Athlon 64, and Athlon XP 3200+ systems.
| Performance Test Configuration | |
| Processor(s): | Intel Pentium 4 at 3.2GHz (800FSB) AMD Athlon64 3200+ (2.0GHz) AMD Athlon64 FX51 (2.2GHz) AMD Athlon XP 3200+ (2.2GHz, 400MHz FSB) |
| RAM: | 2 x 512MB Mushkin PC3500 Level II 2 x 512MB Mushkin ECC Registered PC3200 2 x 256MB Corsair PC3200 TwinX LL (v1.1) |
| Hard Drive(s): | 2X250GB SATA drives in RAID 0 Array Maxtor 120GB 7200 RPM (8MB Buffer) Western Digital 120GB 7200 RPM (8MB Buffer) |
| Video AGP & IDE Bus Master Drivers: | Intel SATA RAID Drivers VIA 4in1 Hyperion 4.49 (August 20, 2003) NVIDIA nForce version 2.45 (7/29/2003) NVIDIA nForce version 2.03 (1/30/03) |
| Video Card(s): | AT Radeon 9800 XT 256MB (AGP 8X) ATI Radeon 9800 PRO 128MB (AGP 8X) |
| Video Drivers: | ATI Catalyst 3.7 |
| Operating System(s): | Windows XP Home (Dell Dimension XPS) Windows XP Professional SP1 |
| Motherboards: | Dell Dimension XPS Gaming System Asus P4C800-E (Intel 875P) 3.2Ghz P4 Abit KV8-MAX3 Gigabyte K8NNXP-940 Athlon64 FX51 Chaintech ZNF3-150 (nForce3) Athlon64 3200+ MSI K8T Neo (VIA K8T800) Athlon64 3200+ DFI NFII Ultra (nForce2 U400) Barton 3200+ |
Recent performance tests on Athlon64, nForce2 Ultra 400 and Intel 875/865 boards used 2 x 512MB Mushkin PC3500 Level II Double-bank memory. The Athlon64 FX requires Registered or Registered ECC memory, so tests with the Gigabyte K8NNXP-940 were performed with Mushkin High Performance Registered ECC DDR400 memory. Previous tests of motherboards used 2 x 256MB Corsair 3200LL Ver. 1.1. Mushkin PC3500 L2 was employed to preserve the 2-2-2-6 timings that were used in tests with Corsair 3200LL Ver. 1.1. Both Mushkin and Corsair use the same Winbond BH5 memory chips in these modules.
All performance tests were run with the ATI 9800 PRO 128MB video card with AGP Aperture set to 128MB with Fast Write enabled. Resolution in all benchmarks is 1024x768x32.
For the fairest comparisons, benchmarks were recompiled for the Asus P4C800-E using a 3.2GHz Pentium4 processor.
Additions to Performance Tests
We have standardized on ZD Labs Internet Content Creation Winstone 2003 and ZD Labs Business Winstone 2002 for system benchmarking.Game Benchmarks
We have added Gun Metal DirectX Benchmark 2 from Yeti Labs, the new X2 Benchmark, which includes Transform and Lighting effects, and Aquamark 3 to our standard game benchmarks. We will be adding other benchmarks in the near future.
The XPS achieved a new high in Content Creation, which has always been one of the better benchmarks for Intel Pentium 4 based machines. With the combination of the 3.2GHz P4, ATI Radon 9800 XT, and the SATA RAID, we see scores approaching 60 for the first time. The Asus P4C800-E should be in the same area, since it was also tested with a 3.2 and 9800 PRO, and the video card does not have a great influence on Multimedia Winstone 2003. The fact that there is still a wide difference between the Asus and Dell in this benchmark tends to indicate that the hard drive has a large influence on the final result, with the huge 500GB SATA RAID pushing the Dell score to new highs. Pentium 4 systems rarely perform as well in General Usage, and the Dell performs here about the same as the Asus P4C800-E system. The newest Athlon64 FX systems dominate this benchmark as we would expect.
You will see in our whole benchmark suite that benchmarks for the Dell XPS are literally all over the place. In some tests, the Dell performs exceptionally well compared to competitive systems and processors, while in other benchmarks, the Dell is inexplicably much poorer than we would expect. We shared our benchmarks with Dell's testing lab before publishing, and in their own test environment, their results were roughly on par with those we found.
Facebook
Twitter
RSS
43 Comments
View All Comments
araczynski - Friday, November 14, 2003 - link
this has got to be one of the most pointless reviews i've ever seen here.this site is for people that are more or less tech-savvy, who like to build their own machines. why would we care about the performance of a pre-packaged-just-add-water box from anybody (at least anybody mainstream). i think its a known fact that you'll never get your money's worth with those.
no, i didn't read the review, and based on teh comments above, i'm glad i didn't waste my time.
please get back to the more meaningfull stuff.
Wesley Fink - Friday, November 14, 2003 - link
#9 - AnandTech works very hard to maintain consistent conditions when we test COMPONENTS, but a system has to be viewed differently. If the system is a barebone it is tested with standard components, but if it is a complete system we do not modify it for testing by removing or changing components. We are not, in this case, testing just a motherboard, or just a drive system, we are testing a complete system as delivered to and used by end users. End users will not remove the Audigy for gaming, not will they replace the 9800 XT with a 9800 PRO to see how games run. We therefore test complete systems, as much as we can, as we receive them. This is particularly true with the Dell XPS, because the user is VERY unlikely to change anything in this machine.We were careful to point out differences from our other benchmarks so readers would not jump to the wrong conclusions. We did consider publishing these benchmarks in isolation, without comparisons, but that would tell you little about the Dell's performance compared to the best components we have tested.
Perhaps you will be more satisfied with the benchmark comparisons when we test a similarly equipped ElitePC system running an Athlon64 FX in an upcoming review.
jlv - Friday, November 14, 2003 - link
This review leaves me with a bad feeling.The XPS won on tests where hard drive speed mattered, and that's because it had a faster disk subsystem (striped SATA). Is that called out in the
"comments"?
Here you are holding up a $3300 system, and it doesn't outperform. (But you give us a "special offer" to buy one ourselves). I'd love to know what the other systems cost, for comparison sake.
Overall, this is the type of "review" that lead me to stop going to TomsHardware.
Possessed Freak - Friday, November 14, 2003 - link
This is probably the only review I have disliked from Anandtech for the past year or two.If you suspect the audigy is causing benches to slow down, and you can't fully disable the device in windows, remove it?
If you have a faster video card than all the other benches, change it?
Offer two bars for the tests, one "as shipped" and one "modified" so we can see if there is something else holding the dell back.
You suspect sata raid is speeding up some benches, as a reviewer isn't it your job to find out? I don't have one of these machines to see just how the raid affects the benches, you do. Run a myriad of tests to find out just why the dell is all over the place in these benchmarks.
If you are not willing to change the dell specs then how hard is it to add audigy's to the other test bed machines, or add the latest vid card?
LoneWolf15 - Friday, November 14, 2003 - link
Quoted from the article:"It is not designed for the hobbyist or overclocker; rather, it is designed and tweaked for the buyers who want their gaming system already set up, already tweaked for best performance, and ready to give great gaming performance right out of the box. As we evaluate the Dell Dimension XPS, this is an important distinction. Overclockers and tweakers will not be satisfied with the XPS"
Guess what? This means that probably 90% or better of Anandtech readers WILL NOT be satisfied with the XPS.
Even if I'm not overclocking, and not pushing my system's BIOS to the limits, I can build a system with similar components far cheaper, with better down-the-road upgradeability and at least equal performance and build quality. I can also do without the LCD --no matter how sleek they look, LCD's still aren't for gamers or video/graphics pros, IMHO.
Dell is fairly good in a business environment (I'm a sysadmin over several hundred of them) though support has gone downhill since they moved their phone centers from the US to Bangalore, India. I'd rather support myself, and I'd think most Anandtech readers would too. I just don't see this system as being targeted at the average Anandtech reader.
AlexWade - Friday, November 14, 2003 - link
Can I gripe about Dell's for a second?Our computer labs just upgraded to Dell's. I swear, they seem so slow. But, here is where my gripe with Dell begins. If I said the LCD has a bad image quality, I'd be giving it too much credit. I can see the individual dots with my face 1 foot away from the screen.
Okay, I'm done complaining.
firtol88 - Friday, November 14, 2003 - link
"Dell shipped the XPS with an ATI Radeon 9800 XT video card, which was used for all benchmarks with the Dell system. Please keep in mind that all of our comparison benchmarks were run with the ATI Radeon 9800 PRO, which is slightly slower."OK so this is really just an add right...
Anyway your conclusion should read 'when equiped with a vidieo card offering an 8% increase in clock speed, a 7% increase in memory speed and double the video memory this Dell XPS system can almost keep up with the other systems we tested...'
TrogdorJW - Friday, November 14, 2003 - link
I think the discrepancies that show up in the testing are very easy to pinpoint. But what a waste of good components. It would be SO easy for Dell to provide a truly great gaming system, but they are held back by their need to be proprietary and not allow any tweaking.First, the defects. As stated, lack of tweaking is not a good thing for most of us. Worse than that, though, is that the system will not even run at the SPD settings in memory modules. I can live without being able to manually set the timings, since this is Dell, but if I have CAS2 2-2-2-6 RAM, by DAMN the system had better run at those settings! That, incidentally, is almost certainly why the system performance drops so far in some tests.
Next, they go with an i875 chipset for the motherboard, but then they don't even give you Intel's Gigabit CSA interface. Is it needed? Not by a lot of people, but come ON, Dell - your in freaking bed with Intel already, so at least give us their top features! For a system targetted at gamers and costing a cool $3G, there's no reason to stick the user with an old 10/100 Ethernet connection.
Now, the case looks pretty cool, for the most part, but that is a butt-ugly cabling job. I realize that IDE cables are spec'ed to be no more than 18 inches long, but if you go with QUALITY cables, you can get longer 24" cables that work fine (especially since they're only for the CD drives), and there *has* to be a way to hide them out of the way a bit better. As much as it pains me to say this (since one of the few companies I hate more than Dell is Apple - I blame it on having to support Dell systems at my job), for a look at how to properly create a fold-open case and still keep the IDE cables tucked away nicely, the G4 systems were great. Again, Dell could have easily done the same, and they SHOULD have done the same for their XPS "Gamer" machine.
While I'm on the subject of the case, I understand that it is overall a pretty well-made system, but the huge cover over the CPU and motherboard just looks fugly if you ask me. Yeah, it works, but it sure ain't gonna win a beauty contest! Also, the 460W power supply is a nice addition. Too bad it's a proprietary Dell design, I'm sure, so if it goes out and your warranty is expired, you'll be spent ordering a replacement from Dell for $200 or so. (Unfortunately, this probably isn't exaggerating!) And WHY IN TARNATION DID THEY PUT THE POWER SUPPLY ON THE BOTTOM!? Oh, wait, there's no tweaking, so you don't need to worry about heat... but *still*, it's just odd. I mean, the PS creates a LOT of heat, heat rises.... Stupid.
Finally, did I mention that Dell systems are PROPRIETARY? It's not a big deal now, when you buy the system and it's all new and happy. But three or four years down the road, when parts start having issues and the warranty is gone, what are you going to do? I've seen proprietary Dell motherboards for old Pentium II systems that will still cost you over $200! And since the power supply, case, and motherboard are all proprietary as well, you're stuck either paying Dell a premium for service, or else you have to buy a new case, power supply, and motherboard should any of these fail after the warranty. DO NOT DISCOUNT THIS COST!!!
Now, with all of my complaints about that out of the way, let me address the benchmarks. Gee... scores seem to be all over the place. Wonder why? Well, my guess is that it's due to bottlenecks that the various benchmarks stress. Common bottlenecks include the memory subsystem (Quake 3 is notorious for stressing this!), hard drives, graphics card, and of course all of these at the same time, in varying degrees.
So it set a new high in Content Creation? No big surprise, since the benchmark is heavily influenced by hard drive speeds, and the striped SATA array is certainly up to snuff. Several of the other benchmarks depend on hard drive speeds to varying degrees, and when this is the case, the Dell gets a boost relative to the other contestants.
How about the low gaming scores and low workstation scores? Check those 3-3-3-8 memory timings out! There is no such thing as "optimizing" a P4 i875 system to run best at 3-3-3-8 timings. The P4 is more forgiving of slower timings than other processors, but lower will ALWAYS be better, or at least no worse. And knowing Dell, the system is probably really running on a 198.5 MHz bus or something, "just to be safe" - basically, the opposite philosophy of Asus and Abit and such.
But what about the areas that the system excels at that CAN'T be attributed to the hard drive array? Well, it DOES have the fastest graphics card around, the 9800XT. That's why some games and some workstation benchmarks score really well. Not all games stress the memory like Quake 3. Aquamark 3 most certainly doesn't care much about memory speeds in comparison to its desire for a fast CPU and graphics card. But a slow memory subsystem will usually hurt games that are highly optimized. (I think most would agree that Quake engine games and Unreal games tend to be very well optimized, whereas 3DMark and Aquamark and some of the other "eye candy" benchmarks aren't written as well, so they stress the CPU or graphics card more.)
Anyway, the overall impression of this system for any but the most hardware impared HAS to be negative. This is a GAMING machine, and yet the only games that it won at, even with it's graphics card advantage, are Aquamark3 and X2. Wait... those aren't even games (yet). Yeah, the 20" LCD display would be pretty nice to have, if you've got the cash. But the only reason to get this over a REAL gaming machine - built by Falcon Northwest, Voodoo, Alienware, or some other company - is that you're too damn chicken to buy anything that isn't made by the same company that provides your "Easy as Hell" work computers.
That's my story, and I'm sticking to it.
Pumpkinierre - Friday, November 14, 2003 - link
Interesting setup- a coupla questions:1)Does Windows XP home ed. support Hyperthreading? (I know it doesnt support SMP)
2)Was there a CPU temperature monitor and if so did you get a squiz at it, at idle and during the benchmarks? I'm interested in that cooling extraction system.
TheInvincibleMustard - Friday, November 14, 2003 - link
And I ask again, is it so hard to somehow highlight the product being tested in the scoring tables? I cannot imagine it's that hard to "paint" that bar graphic red while all the rest are blue ...Aside from that, did Dell offer any reasons for the discrepancies that showed up throughout the testings? Or have you not heard back from them yet?