ECS K7S6A Review, + 1 more from t-break - 74%

[NFS4]

http://www.ocworkbench.com/hardware/elite/k7s6a/k7s6ap1.htm

 

[NFS4]

http://www.tbreak.com/hard/mobo/ecs_k7s6a/page1.html

Poor thing got beat up by a "lowly" KT266A in most of the benches...even at 166MHz memory speeds. Pitiful

 

[kuk]

SiS' ups and downs ... kinda sad.
And ECS shouldn't have dropped the Ethernet port ... which I think is a great addition.

Well, we'll have to wait for KT333 (or even maybe KT333A for the bug fixes) ...

 

[fendel]

Did we read the same review?? Looked to me like this board was neck-and-neck with the KT266A in most of the benchmarks.

 

[NFS4]

<< Did we read the same review?? Looked to me like this board was neck-and-neck with the KT266A in most of the benchmarks. >>


It shouldn't be "neck and neck" It's using DDR333 for goodness sake

 

[StormRider]

Does anyone know if it is possible to use the last PCI slot at the same time as that AMR/CNR slot? Or do they share the same I/O bracket space?

 

[Bovinicus]

Looks like a decent board to me. I would consider it for a low cost system.

 

[formulav8]

<<

<< Did we read the same review?? Looked to me like this board was neck-and-neck with the KT266A in most of the benchmarks. >>


It shouldn't be "neck and neck" It's using DDR333 for goodness sake >>




Its still running on a 133mhz fsb. Everyone knows that the Athlon can't really use anymore bandwidth than what 133mhz provides unless they up the fsb. Just raising the memory bandwidth does little to nothing for the Athlon.

 

[gogeeta13]

It is disapointing for four reasons.
1, the lack of Ethernet
2, the increased price for DDR2700 makes the economics of this board pretty bad, it will be a while till the prices drop and we can truly realize the 1/5 and 1/6 PCI multipliers
3, ECS needs to realize they could gain a lot of respect if they tried a little harder for enthusiasts. While the board may be a success for some OEMs, I will still use the K7S5A for my system builds
4, It scored worse than the k7s5a in I/O and bandwidth stuff..HUH

 

[RagingGuardian]

I wouldn't even consider getting the S6A after all the stability problems I had with the S5A. Of course Pabster will come along any minute to say that I had no idea what I was doing but hey...

 

[Duvie]

So far I am not impressed with the sis745...at least the ECS offering...

I agree the ddr333 seemed disappointing as well as I agree I know we are running up against the bandwidth limitations of the athlon until they make a fsb change, but by going 133/166 the scores shouldn't get worse, period!!!


My sis735 is been a rock( and I am not the only one!!!) and I see nothing in the sis745 review or anything out there to buy that would substantially give me any real world boost in performance...at least that I could feel or see...


Lets see if asus's offerring will do a bit better and give us the use of the firewire support...

 

[Athlon4all]

<< It shouldn't be "neck and neck" It's using DDR333 for goodness sake >>

I think it should've been neck and neck because the Athlon's fsb is allready maxed out by KT266A as it is, so I'm not suprised at all. Really, what I am most concerned with this board is compatibility and stability, the 2 places where I feel K7S5A lacks. We'll see once the board gets into users' hands.

EDIT: I noticed on the BIOS ss that it has 166fsb+ 1/5 divider, I would've liked to see it tested at 166/166, then we might have seen some benefit from PC2700 DDR.

 

[TimeKeeper]

Not that I like to complain nor I have problem w/ this SiS mobo, but the review seems to be little off.

Here is why..................

OCworkbench is using this Kgood4.bin BIOS which has release date of 1/25/02 (10 days after their official S6a10B.bin BIOS)

1. Kgood4.bin boot up w/ "NOT FOR RESALE VERSION".

Not that I have anything against Beta BIOS, but not everyone going to search and download this particular version.

2. Kgood4.bin default 133/133/33 setting has the FSB running at 133.90FSB

The retail S6a10B.bin (dated 1/15/02) at same setting will only run at 132.90FSB
No big different in performance you say? Of course not, but you might not feel happy when you see AMD760 out perform your mobo from Sandra Benchmark or your 1.2Ghz AMD show as 1.19Ghz at your WindowXP "System Property" Menu.

3. CPU Voltage still max out at 1.82V. OCworkbench fail to mention about this part.
NO, my CPU runs fine w/ extra .25V. But I am NOT extreme OC-er.

4. One last weird things.....(perhaps, Sandra 2002 is acting up)
MainBoard Information w/ new BIOS shows that my SIS745 become SiS733 chip set????? HUH??????

 

[NFS4]

<<

<< It shouldn't be "neck and neck" It's using DDR333 for goodness sake >>

I think it should've been neck and neck because the Athlon's fsb is allready maxed out by KT266A as it is, so I'm not suprised at all. Really, what I am most concerned with this board is compatibility and stability, the 2 places where I feel K7S5A lacks. We'll see once the board gets into users' hands.

EDIT: I noticed on the BIOS ss that it has 166fsb+ 1/5 divider, I would've liked to see it tested at 166/166, then we might have seen some benefit from PC2700 DDR. >>


I disagree

If you are going to pay more to be able to use DDR333 memory, you had BETTER be able to see some kind of performance increase. I just ain't seeing it with the SiS745. For a KT266A motherboard with PC2100 to be able to hang with the SiS745 with DDR333 in most benches and exceed it in many others just throws the SiS745 out of my sights. And I was actually thinking about getting a SiS745 based motherboard. And the fact that the numskulls at ECS released the K7S6A w/o an integrated NIC is simply a thumbs down IMHO. ECS is more about getting their sh!t out first than putting together a solid board. That's why they are always the first out of the gate. Everyone else is in house tweaking their boards and ECS is throwing out stuff early. I think that ECS is too early with their releases and Abit is ALWAYS too late. Everyone else falls into "just right"

Guess I'll wait for KT333A to see what it has to offer. Maybe THAT's why we haven't heard Pabster say much about his new K7S6A

He spent all that money on a new motherboard and didn't see any benefits and lost a NIC

 

[Athlon4all]

<< If you are going to pay more to be able to use DDR333 memory, you had BETTER be able to see some kind of performance increase.I just ain't seeing it with the SiS745. For a KT266A motherboard with PC2100 to be able to hang with the SiS745 with DDR333 in most benches and exceed it in many others just throws the SiS745 out of my sights. >>

Well, I am going to point out that when K7S5A first came out, it was $65 and considering that while the overclocking of this board is hardly spectacular, it is something, and it will prolly make it's way down quickly to $55-60 that K7S5A is at now. Frankely, I never once was excited about 745's performance really. I was hoping to see a tad performance increase via the SIS 645's memory controller, but I never once expected it to get a performance boost. For the record, I think that SiS 735 is close enough to 266A in everything but games where it falls behind a tad but it is not big (<5%).

The thing that was big in my book was that at least (at the time) Abit, and Shuttle would be making boards using it and now we've got even more manufacturer's saying that they are developing boards using it. I predict that we will see this exact same picture with KT333(A). Unless T-Bred uses at 166fsb, PC2700 DDR is useless on the Athlon platform, and really, I still maintain that with the right manufacturer, SiS 745 has a lot of potential, but it really, is not that far behind KT266A IMO, and I will not hesitate to recommend boards using SiS 745 if the board is at the same stability and compatibility level as 266A ones, and I still am holding my breath on K7S6A. I guess we'll see if we see similar issues with it like we have with K7S5A.

<< And I was actually thinking about getting a SiS745 based motherboard. >>

Yes, and has that changed? You have made it clear that you will not buy K7S5/6A but what about Abit KS7 or Asus' or MSi's offerings?

<< And the fact that the numskulls at ECS released the K7S6A w/o an integrated NIC is simply a thumbs down IMHO. >>

I agree. Why take it out? That definately was a plus for K7S5A.

 

[NFS4]

<< Yes, and has that changed? You have made it clear that you will not buy K7S5/6A but what about Abit KS7 or Asus' or MSi's offerings? >>


The performance between board manufacturers will not be that different. And even the others are faster than the K7S6A, then it will still only be a little out in front of the KT266A. That just ain't worth it to me. DDR33 + cost of a new motherboard just doesn't justify SiS745's performance right now.

 

[AA0]

I would have liked to see 166 FSB benchmark too, but the problem is both sites couldn't get the board to overclock well at all. Even with mulitipliers unlocked one got it to 138 and the other 150. I see a lot of RMAs coming if their boards can only go to 138, but whats new, its typical of ECS.

 

[Innoka]

"Everyone knows that the Athlon can't really use anymore bandwidth than what 133mhz provides unless they up the fsb"
This comment seems to be expressing that running memory asynchronously with the front side bus produces no gains. The Apollo chipsets showed nearly half the benefit of moving the fsb up to the memory speed as well. It is not trivial for a chipset that can implement it properly. 166 asynchronous over 133 should produce 10% real world gain in bandwidth straining applications.

 

[Innoka]

dp

 

[Budman]

I see no reason for me to get a sis745 chipset now,it's basicaly the same speed as 735.

Looks like I will keep my K7S5A a little while longer,but that's ok mine's running great.

 

[MoleX]

<< I see no reason for me to get a sis745 chipset now,it's basicaly the same speed as 735.

Looks like I will keep my K7S5A a little while longer,but that's ok mine's running great. >>



Ditto

Unless perhaps they change the PCB color

 

[rockhard]

Yeah ECS have really missed the boat this time but not because of performance which is more than adequate for the market sector its aimed at.
The thing that could have made this board great was to make use of all the chipsets capabilities IMO.
All they needed to do was trace and pop a Firewire port on there and the Digital Video guys would have been queing up for it. 5 or 10$ more with integrated Firewire and Lan - why not? Hmmm?
Im hoping Gigabyte will come up with something.
After trying their GA7VTXH im keen to try another of their upcoming boards Be nice if it was a Sis based mobo.

 

[rockhard]

Just been sniffing around OCP and they have some benches showing exactly the same kinda thing is likely to happen with the KT333 chipset and the Athlon/XP etc.
Maybe its AMD you guys guys should be dissing and not necessarily the board manufacturers. After all AMD have held onto the 133 (266DDR) FSB a little too long IMO for it to be healthy, yeah?

 

[Athlon4all]

<< The performance between board manufacturers will not be that different. And even the others are faster than the K7S6A, then it will still only be a little out in front of the KT266A. That just ain't worth it to me. DDR33 + cost of a new motherboard just doesn't justify SiS745's performance right now. >>

Hmmm. I think one thing is that I'm going from the point of view of someone buying new, while you it seems are talking about someoen that allready has 266A am I correct? Let me restate something. Don't get me wrong, I feel that KT266A, and SiS 745 are great chipsets (let's not forget nForce 415-D as well) and I guess to me, I feel that SiS is at, maybe above VIA's level of compatibility, and stability if it is on the right board. In other words, if I had to choose between 2 otherwise identical boards (lets say Abit KS7 and Abit KR7A), I would take the one with the SiS chipset. But, I do agree that unless board makers use the features (especially Firewire), if someone allready has a 266A, then don't bother.

<< Yeah ECS have really missed the boat this time but not because of performance which is more than adequate for the market sector its aimed at.
The thing that could have made this board great was to make use of all the chipsets capabilities IMO.
All they needed to do was trace and pop a Firewire port on there and the Digital Video guys would have been queing up for it. 5 or 10$ more with >>

Totally agree RockHard. That really sucked.

<< Just been sniffing around OCP and they have some benches showing exactly the same kinda thing is likely to happen with the KT333 chipset and the Athlon/XP etc.
Maybe its AMD you guys guys should be dissing and not necessarily the board manufacturers. After all AMD have held onto the 133 (266DDR) FSB a little too long IMO for it to be healthy, yeah? >>

Where? I don't feel like digging. Could you give a link? Btw, you need to type HardOCP, [ H ]ardOCP or you highlite the text

<< This comment seems to be expressing that running memory asynchronously with the front side bus produces no gains. The Apollo chipsets showed nearly half the benefit of moving the fsb up to the memory speed as well. It is not trivial for a chipset that can implement it properly. 166 asynchronous over 133 should produce 10% real world gain in bandwidth straining applications. >>

When you say Apollo you mean KT133 and KX133? You are correct that running 200fsb Athlon's with PC133 was benefical, but you need to understand how the situation has changed since KT133.
Back in the early Thunderbird days, there was no DDR BUT, the front side bus of the Thunderbird was DDR. The fsb of the T-Bird was 100MHz DDR (200MHz effective) and provided 1.6GB/ps of bandwidth, and when the memory bus was kept in sync with the fsb, the memory bus would be running at PC100 which provides 800MB/ps while the fsb basically was providing double that bandwidth. As a result, KT133's PC133 support while the fsb was running at 100MHz DDR did provide a noticeable increase in performance because there was still bandwidth on the fsb that wasn't being used. Not even PC133 could saturate the 100MHz DDR fsb (PC133 provides 1.064GB/ps of bandy, while the fsb still was providing 1.6GB/ps. Still, 500MB was left to be used). Now, when DDR came upon the scene, PC1600 DDR provided 1.6GB/ps just like the 200MHz fsb and with 266fsb CPU's, same thing as long as PC2100 DDR was being used. Now, the reason, PC2700 doesn't on SiS 745 and won't help on KT333(A) is because, unlike the situation KT133 came into, KT333(A)/745 comes in with faster memory, but the fsb is allready being saturated by PC2100, and thus the extra bandwidth on the memory bus can't be used.

The thing with KT133 applies to Sis 645 for the Pentium 4 as well. The Pentium 4's fsb is 100MHz QDR (providing 3.2GB/ps of bandwidth), and lets take a synchorous configuration for 645 which isn't very common (PC1600). Again, PC1600 provides 1.6GB/ps, while the fsb still has an extra 1.6GB/ps that can't be used, and this is why 645 gets such a benefit from using PC2100 or even PC2700 is because the fsb still isn't even being maxed out. Do you see what I'm saying now?

 

[nortexoid]

blah blah blah blah

for ~$70 the board is a killer deal....also, ddr333 is a bunch of garbage, especially if you're running it asynchronously on any chipset....don't expect a jump in performance w/ the added latency...i don't even know why they'd even offer asynchronous settings anymore; even when they do increase performance, it's negligible.

the SIS 745 is an excellent chipset - more so than the 735....that is, if manufacturers put the integrated firewire to use, otherwise it's virtually identical...
the single chip chipset has everything you'd want, and the performance isn't too shabby either.

board looks good to me - again, i agree w/ the rest that they should've kept the onboard LAN - idiots.

I'd pick this board up without regrets in a second.