Foxconn and Gigabyte Tackle Socket AM2
by Jarred Walton on June 22, 2006 1:30 AM EST- Posted in
- Motherboards
Motherboard Features
We'll begin with a brief overview of the features before we move into the individual motherboard reviews. The following table offers a quick breakdown of the features provided.
Both motherboards use the same chipset, so they both have a Northbridge and Southbridge. They also support X16 bandwidth (4 GBps bidirectional) to both X16 PCI-E slots while less expensive boards based off the nForce 570 SLI chipset have to "make do" with X8 bandwidth when running in SLI/multi-GPU mode. In actual practice, this usually doesn't make more than a slight difference in gaming performance, especially when you consider the fact that the HyperTransport link to the processor only provides enough bandwidth (4 GBps bidirectional) to max out a single X16 connection. We have seen in the past that the "true dual X16" solutions are often a bit faster at high resolutions and detail settings, but the differences are just as likely to come from better BIOS tuning for games as they are from the dual X16 PCI-E implementation.
One of the interesting differences is that the Gigabyte board comes with three X16 slots, although the middle X16 slot only offers X8 bandwidth. As PCI Express cards are downwards and upwards compatible -- i.e., you can run X1, X2, X4, X8, and X16 cards in a physical X16 slot, and you can also use an X16 slot with everything from X1 through X16 data connections -- the inclusion of the third X16 slot could prove useful in the future. ATI has recently talked about the ability to use one of their graphics cards as a physics processing device, so potentially you could move older graphics hardware into the extra slot when you upgrade. Where the Gigabyte board has an extra X16 slot, the Foxconn has an X4 slot, and the Gigabyte board also has an additional X1 slot above the top PEG slot. Using the extra X1 slot may prove somewhat difficult, however, as the large heatsink on the Northbridge will interfere with anything other than a very short X1 card.
There are a few other differences in terms of features. The Gigabyte board is entirely passively cooled, while the Foxconn board has a fan on the Southbridge. (We'll have a bit more to say about the cooling solutions when we get into the specifics of each board.) The Gigabyte board also uses heatpipes connecting the Northbridge to the Southbridge, as well as to a third radiator/heatsink that helps to cool the MOSFETs. Both boards offer six SATA ports provided by the Southbridge, but the Gigabyte board adds a J-Micron chip to provide two additional SATA ports. (All of the SATA ports support 3.0Gbps transfer rates.) The SATA ports on the Gigabyte board also use the new clamp and latch design, and the included SATA cables also support the future; the Foxconn board opts for the more traditional style. Finally, the Gigabyte board has Gigabyte's Dual BIOS feature, which allows recovery from failed flash attempts. The Foxconn board does have a couple extras, though: it has an onboard debug/POST code LED, and it includes onboard reset and power buttons.
Overall, we have to give a slight advantage to the Gigabyte board in terms of features, but the Foxconn board is no slouch either. The deciding factor for many people is going to come down to price and availability, so for the time being the Foxconn board has the lead while we wait for the Gigabyte board to begin shipping. There are also competing offerings from other manufacturers, so we'll hold off declaring a victor until we finish our initial AM2 motherboard reviews. Now let's get to the finer details of these two boards.
We'll begin with a brief overview of the features before we move into the individual motherboard reviews. The following table offers a quick breakdown of the features provided.
| Motherboard Specifications | ||
| Foxconn C51-XEM2AA | Gigabyte GA-M59SLI-S5 | |
| Market Segment | Enthusiast - SLI (2x16) | Enthusiast - SLI (2x16) |
| CPU Interface | AM2 | AM2 |
| SLI Technology | Yes | Yes |
| NVIDIA LinkBoost | Yes | Yes |
| NVIDIA FirstPacket | Yes | Yes |
| NVIDIA DualNet | Yes | Yes |
| Memory - EPP | Yes | Yes |
| Teaming | Yes | Yes |
| TCP/IP Acceleration | Yes | Yes |
| MediaShield | Yes | Yes |
| GPU Ex | Yes | Yes |
| Chipset | nForce 590 SLI | nForce 590 SLI |
| MCP Voltage | Auto, 1.525V ~ 1.700V in .025V increments | Normal, + .1V, .2V, .3V |
| Memory Speeds | DDR2 - Auto, 400, 533, 667, 800 | DDR2 - Auto, 400, 533, 667, 800 |
| Memory Voltage | 1.825V ~ 2.500V, .025V increments | Normal, + .1V ~ .7V in .1V increments |
| Memory Settings | tCL, tRAS, tRP, tRCD, tRPD, tRC, CMD, tWR, tRWT, tWTR, tREF, tRRD, DQS Skews, Async Latencies, 7 Drive Strength settings |
tCL, tRAS, tRP, tRCD, tRPD, tRC, CMD, tWR, tRWT, tWTR, tREF |
| PCI Express Speeds | 100MHz~200MHz | 100MHz~200MHz |
| HyperTransport Frequency | 100MHz~500MHz in various increments | 100MHz~500MHz in various increments |
| HyperTransport Multiplier | Auto, 1x, 2x, 3x, 4x, 5x AM2 to NB, NB to SB |
Auto, 1x, 2x, 3x, 4x, 5x AM2 to NB, NB to SB |
| HyperTransport Link Width | 8/8, 16/16 - NB and SB | 8/8, 16/16 - NB and SB |
| HyperTransport Link Voltage | Auto, 1.325V ~ 1.500V in .025V increments | Auto, 1.25V, 1.3V, 1.35V, 1.4V |
| CPU Clock Multiplier | Auto, 4x ~ 12x (4800+) Maximum Dependent upon CPU | Auto, 4x ~ 12x (4800+) Maximum Dependent upon CPU |
| CPU Voltage | Auto, .3750V to 1.8500V in .025V or .0125V increments | Normal, .800V to 1.550V in .025 increments |
| Memory Slots | 4 x 240-pin DIMM Slots 8GB Capacity |
4 x 240-pin DIMM Slots 16GB Capacity |
| Expansion Slots | 2 x PCI Express X16 1 x PCI Express X4 1x PCI Express X1 2 x PCI 2.3 |
2 x PCI Express X16 1 x PCI Express X8 2 x PCI Express X1 2 x PCI 2.3 |
| Onboard SATA 2.0 | 6 x SATA 3Gb/s Ports - NVIDIA | 6 x SATA 3Gb/s Ports - NVIDIA 2 x 3Gb/s Ports - JMicron B363 |
| Onboard SATA 2.0 RAID | RAID 0, 0+1, 5, JBOD | RAID 0, 0+1, 5, JBOD |
| Onboard IDE | 1 x UltraDMA Connector 133/10/66/33, Two Drive Support | 1 x UltraDMA Connector 133/10/66/33, Two Drive Support |
| Onboard USB2.0 | Six via Rear Panel, Four via motherboard headers | Four via Rear Panel, Six via motherboard headers |
| Onboard IEEE-1394a | TI TSB82AA2 1394a + TI TSB81BA3 1394b |
TI TSB43AB23 1394a |
| Onboard LAN | 2 GbE - Marvell 88E1121 | 2 GbE - Marvell 88E1116 |
| Onboard Audio | Realtek ALC882D | Realtek ALC888 |
| Power Connectors | 8-pin ATX, 24-Pin ATX, 4-pin Molex | 8-pin ATX, 24-Pin ATX, 4-pin Molex |
| Back Panel I/O Ports | 1 x PS/2 Keyboard 1 x PS/2 Mouse 1 x Audio I/O Panel 2 x RJ45 LAN 6 x USB 2.0 1 x IEEE-1394a and 1394b |
1 x PS/2 Keyboard 1 x PS/2 Mouse 1 x Audio I/O Panel 2 x RJ45 LAN 4 x USB 2.0 1 x Parallel Port 1 x Serial Port 1 x S/PDIF Optical Out 1 x IEEE-1394a |
| NTune 5.0 Support | Complete | Limited |
| BIOS | Award, P20 | Award, F2 |
Both motherboards use the same chipset, so they both have a Northbridge and Southbridge. They also support X16 bandwidth (4 GBps bidirectional) to both X16 PCI-E slots while less expensive boards based off the nForce 570 SLI chipset have to "make do" with X8 bandwidth when running in SLI/multi-GPU mode. In actual practice, this usually doesn't make more than a slight difference in gaming performance, especially when you consider the fact that the HyperTransport link to the processor only provides enough bandwidth (4 GBps bidirectional) to max out a single X16 connection. We have seen in the past that the "true dual X16" solutions are often a bit faster at high resolutions and detail settings, but the differences are just as likely to come from better BIOS tuning for games as they are from the dual X16 PCI-E implementation.
One of the interesting differences is that the Gigabyte board comes with three X16 slots, although the middle X16 slot only offers X8 bandwidth. As PCI Express cards are downwards and upwards compatible -- i.e., you can run X1, X2, X4, X8, and X16 cards in a physical X16 slot, and you can also use an X16 slot with everything from X1 through X16 data connections -- the inclusion of the third X16 slot could prove useful in the future. ATI has recently talked about the ability to use one of their graphics cards as a physics processing device, so potentially you could move older graphics hardware into the extra slot when you upgrade. Where the Gigabyte board has an extra X16 slot, the Foxconn has an X4 slot, and the Gigabyte board also has an additional X1 slot above the top PEG slot. Using the extra X1 slot may prove somewhat difficult, however, as the large heatsink on the Northbridge will interfere with anything other than a very short X1 card.
There are a few other differences in terms of features. The Gigabyte board is entirely passively cooled, while the Foxconn board has a fan on the Southbridge. (We'll have a bit more to say about the cooling solutions when we get into the specifics of each board.) The Gigabyte board also uses heatpipes connecting the Northbridge to the Southbridge, as well as to a third radiator/heatsink that helps to cool the MOSFETs. Both boards offer six SATA ports provided by the Southbridge, but the Gigabyte board adds a J-Micron chip to provide two additional SATA ports. (All of the SATA ports support 3.0Gbps transfer rates.) The SATA ports on the Gigabyte board also use the new clamp and latch design, and the included SATA cables also support the future; the Foxconn board opts for the more traditional style. Finally, the Gigabyte board has Gigabyte's Dual BIOS feature, which allows recovery from failed flash attempts. The Foxconn board does have a couple extras, though: it has an onboard debug/POST code LED, and it includes onboard reset and power buttons.
Overall, we have to give a slight advantage to the Gigabyte board in terms of features, but the Foxconn board is no slouch either. The deciding factor for many people is going to come down to price and availability, so for the time being the Foxconn board has the lead while we wait for the Gigabyte board to begin shipping. There are also competing offerings from other manufacturers, so we'll hold off declaring a victor until we finish our initial AM2 motherboard reviews. Now let's get to the finer details of these two boards.
Facebook
Twitter
RSS
36 Comments
View All Comments
Kougar - Friday, June 23, 2006 - link
Just wanted to say thanks for another good review!Really appreciated all the charts and info on the sound, as I've been eyeing the $25 Audigy2 SE (SB570) card that's now out as an upgrade to my old AC97, since it appears to be better than the older Audigy2 ZS (SB360) in some areas... Primary motivation was to offload work from the CPU. Hmm
Operandi - Thursday, June 22, 2006 - link
I think we've gotten to the point now where performing benchmarks on these boards has gotten to the point to where it should be of little concern. I/O performance is slightly more useful but again generally the same within the same chipset family.I not saying the benchmarks are not important, they are. But from a real world stand point it’s so close you’d never now the difference if somebody didn’t illustrate it for you in a pretty graph.
I would like to see more attention paid to board features, by that I mean features aside from overclocking options. Some of us actually have other priorities other then all out performance. For awhile there you guys were looking at fan control for example. What happened to that?, that’s something I found useful, loud fan’s are no fun.
It would also be nice to know about the bundle of the board, what it includes and what it doesn’t, w/ pics.
Providing some information some information on the caps used on the board would also be appreciated.
JarredWalton - Thursday, June 22, 2006 - link
I did try to focus more on features as opposed to benchmarks. The benchmarks are mostly used to show that none of the boards have severe performance problems. Overall, I think the Gigabyte board has more features, with the exception of FireWire 1394b support. I would also say that right now the Foxconn motherboard has the better BIOS implementation.In terms of fan control, both motherboards reviewed here do have "smart fan speed" options in the BIOS. Given that there are no fans on the Gigabyte motherboard, noise is something that will be created by the CPU fan, GPU fan(s), and case fans rather than the motherboard. If you are worried about system noise levels, the choice of cooling and case is going to be almost certainly be more important than the motherboard.
Capacitors? I know the Foxconn board uses some Rubycon capacitors; you can see the label on the close-up shot. My understanding is that these are good capacitors, but honestly I don't know. The Gigabyte board uses Rubycon capacitors as well, though it's a bit more difficult to see the name in the enlarged image of this area.
As far as accessories goes, we only received a preproduction motherboard from Gigabyte, with a few accessories. We are waiting for the complete retail board to show up, and I will provide an image here once that arrives. I did take a picture of the Foxconn accessories, though I didn't include it in the actual article because it just didn't seem that important. For those that are interested, here it is:
http://images.anandtech.com/reviews/motherboards/r...">Foxconn Accessories
Operandi - Friday, June 23, 2006 - link
Since a system is only as quiet as its loudest fan, fan control is a very important tool. In my opinion just as important as heatsink, fan, and case choice.The whole idea behind BIOS (or software) level fan control is that it offers dynamic control over all of those loud CPU, system, as well northbridge fans; it has very little to do with the noise output of the motherboard itself. Basically it’s much better alternative to manual fan controllers, you get the cooling performance when you need it (gaming, encoding, ect.) and much quieter system then would otherwise be possible at idle. Having built several high-end yet dead silent systems it’s pretty much a requirement for me at this point.
More to the point fan control is a feature and since we both agree that features are more important then the benchmarks why exclude it?
JarredWalton - Saturday, June 24, 2006 - link
I can't remember the last time I encountered a BIOS without fan speed controls. It wasn't excluded because it's not important, but more because it's so common now that I take it for granted. Sorry. Will try to make sure I mention this next time. :)Operandi - Tuesday, June 27, 2006 - link
Excellent ;)BIOS fan control is kind of like overclocking options. Most boards at least support it but some are obviously better then others.
Example, some boards may only control the CPU fan, or offer very little in the way of flexibility. Others are much better; and can control multiple fans and offer a wide range of temperatures and speed settings.
Per Hansson - Friday, June 23, 2006 - link
The caps used on the Foxconn are Rubycon MCZ which is the best cap out there period, however I also saw a small Ost cap in the high-res picture but generally caps below 1000µF rarley fail (and Ost is among the better of the crap brands...)And to answer the comment about Rubycon and Taiwan; They make most of their caps in Japan, check their site out, Rubycon is one of the largest and definantley best regarded capacitor manufacturer. AFAIK they only have a branch office in Taiwan
I think mainboard manufacturers that only use good capacitor brands should deserve a definitive extra point for this, because unlike the other comment the "capacitor issue" is not at all over, it is only getting worse with increased demands on newer processors with more cores and higher power usage... Currently there are more Pentium 4 boards with Badcaps than there are AMD being reported on the site Badcaps.net but that is most likley because the P4 boards simply demand more from the capacitors so they fail sooner. If crap caps are used on AMD boards it is however just a matter of time untill they will also fail.
There is a reason Abit took the plunge after the lawsuit against them and went with 100% Rubycon only capacitors...
Just FYI on the Foxconn I spotted these cap brands:
Rubycon MCZ for VRM input (12v side)
Fujitsu fpcap (V-Core side of VRM, the yellow caps...)
United Chemi Con (unknown model but probably KZG, spread around the board...)
On other samples I have also seen Teapo and Lelon, the Lelon are utter crap, in the same level as GSC, on this sample there where not Lelon though, easy to spot with their light green color...
So the Foxconn is IMO a mixed bag, on one side they really spent some money for good caps around the VRM and also the nice Chemicon caps spread around the board, but cheaped out on the TEAPO, Ost and other misc stuff... The bad caps might be ok though since they are below 1000µF but might still cause the strange mysterious issues like power down/up failures (cold boot issues) etc...
The Gigabyte uses 2x 3300µF Sanyo (probably WG) caps for CPU VCORE, along with 4x Sanyo solid Polymers, a very odd combination that I have not seen used elsewhere... Then it uses only 3x input caps for the 4-phase VRM however they look pretty large so they might hold more capacitance than more regular layouts with one cap for each phase... (they appear to be Nichicon but I'm not certain)
Next there are a few United Chemicon caps spread out, along with Nichicon HM
Had this been an older board the Nichicon HM would have been given a warning but that issue seems to have been fixed by now (there was a bad batch that affected even Intel mainboards...) Still if they are old capacitors from stock there could be problems, but I think Gigabyte known better than this...
So to recap (haha) I would say that the Gigabyte gets my 2x thumps up for Capacitor choices and the Foxconn only gets one due to cheaping out on the small capacitance caps, of course with the note that I own neither board and am only going by the not so high resolution pictures of this review...
But all in all a heads up to Anandtech for providing high-res pictures of some of the capacitors, good job
sprockkets - Friday, June 23, 2006 - link
Just to add, yes, Lelon caps exploded on my MSI boards, and on my Chaintech 7SID0 board, a great SIS735 board, the OST caps are exploding on them. Oddly enough though, the OST caps on my old Shuttle XPC SS40G are not.On the Foxconn board, their uATX nforce 4 board, it also uses rubycon for the power supply side but cheap ones for the rest.
JarredWalton - Friday, June 23, 2006 - link
Damn... I humbly bow to your *far* superior capacitor knowledge. I did get hit by the badcap bug with an older Abit BE6-II board, and had a few systems I built fail for similar reasons. Thankfully, no recent problems for me. I've tried to get better pictures of the capacitors, but I have yet to figure out how to get clear pics without depth-of-field coming into play. Gary managed to get much better results, so most of the images are from him.BIOS failures are my problem lately... I've had four boards die during BIOS updates the past 6 months, and two last week. (Possibly an early AM2 issue).
Per Hansson - Friday, June 23, 2006 - link
Well, thank you but it's become more of a necessity to learn (by replacing dead capacitors on mobos because the mobo manufacturers cheaped out in the first place...)About the pics, I assume the Sony DSC-H1 camera is Garys? Quite impressive shots, I've been looking for a replacement for my Konica KD-400Z but never found anything with a much better macro function than my Konica for the use I have... And don't wanna go SLR...