Exclusive: SLI Head-to-Head: Monarch Micro-ATX vs. Shuttle SFF
by Jarred Walton on March 6, 2006 8:56 AM EST- Posted in
- Systems
Benchmarks Setup
We've talked about the two systems quite a bit, but now, we come down to what many may feel is the most important aspect: which system is faster? It's not entirely fair to compare the Revenge with an SN26P running lower spec graphics cards. However, it's not entirely unfair to make such a comparison either - that's the whole point in getting a uATX case instead of SFF case. We ended up running quite a few different test configurations through our benchmarks in order to provide a complete performance picture. Here are the specific configurations tested.
First, we have the apples-to-apples comparisons. The Shuttle case couldn't hold two GTX KO cards, and I didn't have two regular GTX cards available, so you can compare the single 7800 GTX KO results, both with the FX-57 and 2GB of RAM as well as with a 3800+ and 1GB of RAM. I also ran two 6800 GS cards in SLI and single card mode on both systems, with an Athlon 64 3800+ and 2x512MB RAM. That provides four points of comparison, though the Hornet still has the advantage of running a 150GB Raptor vs. a 250GB Seagate 7200.8. (With only one 150GB Raptor on hand, benchmarking time would have doubled to get a completely equal comparison.)
Besides the "equal footing" tests, I ran some additional configurations through the benchmarks. These additional tests are not meant to be compared directly between Monarch and Shuttle systems, but they are there to help give a better overall picture of performance. I ran the SN26P with the 6800GS cards and the FX-57 processor with 2GB of RAM, to show if and when those cards might still be CPU-limited and where they're GPU-limited. Similarly, I ran the Hornet with the GTX KO cards and the 3800+ processor and 1GB of RAM, so you can see just how much performance you lose - or don't as the case may be - by cutting CPU and RAM costs. There's also the top-end Revenge configuration in its default setup, running the FX-57, 2GB RAM, and the SLI'ed GTX KO cards. Time constraints didn't allow for further detailed testing, but this should be enough information to reach some reasonable conclusions.
The graphs are colored to help compare configurations. For the non-gaming benchmarks, the apples-to-apples results are all the same color. Red is used for the single GTX KO with FX-57; gold is used for 6800 GS SLI, green for a single 6800 GS, and orange for the GTX KO with 3800+. The remaining results are in blue. For the gaming benchmarks, resolutions from 800x600 through 1600x1200 were tested, with and without anti-aliasing. (Anisotropic filtering was always set to 8X.) The gaming results are separated into graphs for 6800 GS configurations and 7800 GTX KO configurations - there's already a lot of overlap, and putting all 11 results in the same chart made it even more difficult to read.
We've talked about the two systems quite a bit, but now, we come down to what many may feel is the most important aspect: which system is faster? It's not entirely fair to compare the Revenge with an SN26P running lower spec graphics cards. However, it's not entirely unfair to make such a comparison either - that's the whole point in getting a uATX case instead of SFF case. We ended up running quite a few different test configurations through our benchmarks in order to provide a complete performance picture. Here are the specific configurations tested.
| Monarch Hornet Pro Revenge Special Edition | |
| Motherboard: | EVGA 131-K8-NF44-AX NF4 SLI (NVIDIA nForce4 SLI chipset) Micro-ATX |
| Processors: | AMD Athlon 64 FX-57 AMD Athlon 64 3800+ (Newcastle) |
| Heatsink: | Thermaltake A1770 |
| RAM: | 2x1024MB PC-3200 OCZ Platinum 2-3-2-7-1T (OCZ4002048ELDCPE-K) 2x512MB PC-6400 OCZ Platinum 2-2-2-7-1T (OCZ600512EEPE) |
| Graphics Cards: | 1 or 2 x EVGA GeForce 7800 GTX KO 256MB 1 or 2 x EVGA 6800 GS 256MB |
| Hard Drive: | Western Digital 150 GB 16MB Cache 10K Raptor (WD1500ADFD) |
| Optical Drive: | Sony 16x DVD+/-RW |
| Floppy Drive: | Mitsumi 3.5 Floppy plus 7-in-1 USB Card Reader |
| Audio: | RealTek AC'97 7.1 (ALC850) |
| Power Suply: | Topower TOP350FX Micro ATX 350W (400W Peak) Note: Shipping units will use a TOP400FX uATX 400W (450W Peak) PSU |
| Drivers: | NVIDIA nForce4 6.70; NVIDIA ForceWare 81.98 |
| Operating System: | Windows XP Professional SP2 |
| Shuttle XPC SN26P | |
| Motherboard: | Shuttle SN26 (NVIDIA nForce4 SLI chipset) - proprietary form factor |
| Processors: | AMD Athlon 64 FX-57 AMD Athlon 64 3800+ (Newcastle) |
| Heatsink: | ICE (Integrated Cooling Engine) 4G |
| RAM: | 2x1024MB PC-3200 OCZ Platinum 2-3-2-7-1T (OCZ4002048ELDCPE-K) 2x512MB PC-6400 OCZ Platinum 2-2-2-7-1T (OCZ600512EEPE) |
| Graphics Cards: | 1 x EVGA GeForce 7800 GTX KO 256MB 1 or 2 x EVGA 6800 GS 256MB |
| Hard Drive: | Western Digital 150 GB 16MB Cache 10K Raptor (WD1500ADFD) |
| Optical Drive: | BenQ 16X DVD-ROM |
| Audio: | VIA 24-bit Vinyl Audio 7.1 |
| Power Suply: | Silent X 350W |
| Drivers: | NVIDIA nForce4 6.70; NVIDIA ForceWare 81.98 |
| Operating System: | Windows XP Professional SP2 |
First, we have the apples-to-apples comparisons. The Shuttle case couldn't hold two GTX KO cards, and I didn't have two regular GTX cards available, so you can compare the single 7800 GTX KO results, both with the FX-57 and 2GB of RAM as well as with a 3800+ and 1GB of RAM. I also ran two 6800 GS cards in SLI and single card mode on both systems, with an Athlon 64 3800+ and 2x512MB RAM. That provides four points of comparison, though the Hornet still has the advantage of running a 150GB Raptor vs. a 250GB Seagate 7200.8. (With only one 150GB Raptor on hand, benchmarking time would have doubled to get a completely equal comparison.)
Besides the "equal footing" tests, I ran some additional configurations through the benchmarks. These additional tests are not meant to be compared directly between Monarch and Shuttle systems, but they are there to help give a better overall picture of performance. I ran the SN26P with the 6800GS cards and the FX-57 processor with 2GB of RAM, to show if and when those cards might still be CPU-limited and where they're GPU-limited. Similarly, I ran the Hornet with the GTX KO cards and the 3800+ processor and 1GB of RAM, so you can see just how much performance you lose - or don't as the case may be - by cutting CPU and RAM costs. There's also the top-end Revenge configuration in its default setup, running the FX-57, 2GB RAM, and the SLI'ed GTX KO cards. Time constraints didn't allow for further detailed testing, but this should be enough information to reach some reasonable conclusions.
The graphs are colored to help compare configurations. For the non-gaming benchmarks, the apples-to-apples results are all the same color. Red is used for the single GTX KO with FX-57; gold is used for 6800 GS SLI, green for a single 6800 GS, and orange for the GTX KO with 3800+. The remaining results are in blue. For the gaming benchmarks, resolutions from 800x600 through 1600x1200 were tested, with and without anti-aliasing. (Anisotropic filtering was always set to 8X.) The gaming results are separated into graphs for 6800 GS configurations and 7800 GTX KO configurations - there's already a lot of overlap, and putting all 11 results in the same chart made it even more difficult to read.
Facebook
Twitter
RSS
13 Comments
View All Comments
JarredWalton - Monday, March 6, 2006 - link
HDCP support is a graphics/display issue. As has been reported, HDCP is not supported on any current retail graphics cards. It's also not supported under Windows XP. We should start seeing HDCP enabled cards (meaning, with the necessary decryption chip) in the near future. The GPUs are ready, but they still need the appropriate chip soldered onto the boards.Personally, I'm really not happy with HDCP at all, so I'm doing my best to avoid it. 1280x720 DivX looks quite nice and runs flawlessly on current hardware. Here's an example from the olympics (18GB compressed to 4.5GB 1280x720):
http://images.anandtech.com/reviews/multimedia/tvt...">2006 Olympics Men's Hockey Gold Match
AGAC - Tuesday, March 7, 2006 - link
Hey, what's to love about HDCP. That said, it seems that we just will have to swallow that frog... I mean, DivX does look nice indeed. The problem is availability of mainstream content. I think it's going to be a very cold day in hell before you can walk in the regular video rental and get the latest blockbuster title in beautiful DivX 1280x720.DHCP will be broken, we all know that. It only harms the legal user because one will have to upgrade video cards, monitors and god knows what more will not be HDCP compliant. Thanks for the your tip and simpathy. Keep up the good work.
AGAC
DigitalFreak - Monday, March 6, 2006 - link
NT