Slashdot Mirror
AMD 760MP Reviews Galore
Keith Whitsitt writes: "Well the NDA seems to be up on AMD's 760MP chipset, and several hardware sites have a review up. So far Anandtech, 2CPU, SimHQ, and Accelnation all have reviews up of this beast. It sure does look like the 760MP has shaped up to be all we expected it to be and more." Time-on-target hype.
But I've not even seen a rumor of an ABIT/ASUS/FIC/Iwill board yet.
Well, no time like the present to start one, then. Psst, word on the street is that ABIT and ASUS are working on ATX-sized 760MP boards.
it has AGP Pro 110 support, this chipset is as much or more for high end PC workstations than low end servers.
Yabbut - what if you already have a bunch of perfectly good high quality 500W+ 100% ATX-compliant power supplies. Tyan has not yet, to my knowledge provided an sufficent explanation of the non-standard connectors on the board. Given that the PS they have been supplying to reviewers are in the low 400W range, and that 600W and larger standard ATX PSs are available, I have a hard time believing that delivering more power to the board is the reason.
Let's see... car payment, or dual Athlons (which are $250 each...)
- A.P.
--
Forget Napster. Why not really break the law?
"Remember when the U.S. had a drug problem, and then we declared a War On Drugs, and now you can't buy drugs anymore?"
What ever happened to Micron's SMP Athlon chipset, the one with the 8MB L3 cache on the chipset die? A flury of coverage a while back, and then nothing (I've looked).
Plato seems wrong to me today
One of the things that I've really liked about Linux is that it seems to "stretch" old hardware, in so far that you don't need to buy the "processor of the week" in order to run the latest and greatest software.
My P1-233MMX has given yeoman service since I got it in 1997, and it has handled everything I've thrown at it just fine.
But now that I've started doing some video editing, som CAD, and some other heavy lifting on it, it's started feeling "slow", so I've been keeping my eye out for something I can pick up that will last another four years or so. It sounds to me like Athlon SMP is just that. More computing HP than I can possibly use at the moment, and at a price that is comparable to what I paid for my Pentium system in '97.
AMD, you'll be gettting my ducets.
DG
Want to learn about race cars? Read my Book
What's going wrong with DMA on your system? I had an S1590 last year, and had UDMA/33 working fine. I remember having to patch the (2.2) kernel originally, but I thought MVP3 support got folded into the later 2.2 versions. Aside from turning on MVP3 support in the kernel and running "hdparm --make-go-fast" there wasn't anything to it.
What made me ditch the board was it's AGP slot that refused to supply AGP level power. Worked fine with a Millenium II, but crashed every two minutes (even in DOS or at the Linux console) with a GeForce 256.
This Tyan board is clearly targeted at the real high-end of the market. I mean 64 bit PCI slots, *dual channel* U160 card, dual port 3com 10/100 card, .... sounds yummy if you're building a big server. But what I'm looking for is a lower-end board, without the built-in SCSI, LAN, video, etc. so that I can build a really affordable SMP system for myself. Any info on when these things are coming?
___
___
If you think big enough, you'll never have to do it.
Your little "summary" seems to imply that *all* of the above tests are somehow included in the overall system performance. They are not. The overall system performance links to the SYSMark test which has nothing to do with any of the above tests. It is correct that Xeon does better on SYSMark because of Content Creation part of this benchmark (probably because of SSE optimizations and/or greater bandwidth of Rambus). But it is misleading to imply that 2 Athlon are slower than 2 Xeons overall. In fact, the benchmarks show the opposite picture.
___
___
If you think big enough, you'll never have to do it.
That said, it's still a pain that you have to have a new PS for the MB.
--
That said, I'm glad to see extra competition in the marketplace; CPU power has ramped up considerably since the Athlon debued and gave Intel a scare.
Also, I'll probably end up buying a 760MP fairly soon :)
--
Yabbut yer building a server when you buy one of these mobos. So you're gonna be willing to pay some coin for a big, reliable power supply -- or, probably, two of 'em, so you have redundancy.
Hardware like this, you don't cheap out on the components!
--
--
Don't like it? Respond with words, not karma.
"The U.S. Environmental Protection Agency and the U.S. Department of Energy awarded AMD an ENERGY STAR® Certificate of Recognition for its energy-efficient processors, including the AMD Athlon MP processor. These processors help manufacturers meet stringent ENERGY STAR specifications for a variety of appliances, equipment and other products. Products with the ENERGY STAR label are designed to use less energy, save money and help protect the environment."
From the pressrelase.
It's called new wave but it's just the same.
Sure, the BP6 was for Celerons, but it was a major hit amongst hardcore overclockers and hardware junkies.
What would you say that the odds are that Abit has a "BP6" in development for the Palimino chips?
Personally I would say that it is very good. Sure, it won't have the on board SCSI, or LAN or Video etc... It won't cost as much however, and knowing the softbios features of Abit, it will be very overclockable.
In all honesty I have heard of NOTHING about a dual AMD board from Abit, but I would put money on one being in the wings.
If a "BP6" for the Palimino comes out, you can be certain that it will be the board to have. Here is hoping...
(Has anyone heard anything about dual AMD boards from other manufacturers other than Tyan?)
Try to hack my 31337 firewall!
This box is an Athlon TBird-800, with an Abit KT7-RAID mobo (VIA KT133 chipset), and the only problem I have is an occasional hardlock on startup (not more than once every 15 boots)
Other hardware: GF2 GTS, SBLive!. Busmastering PCI works fine, and IDE...no problems, except for the shoddy WD drive that I'm replacing as soon as possible.
(Yes, this box runs windows. Just a disclaimer)
-- Veni, vidi, dormivi
but I can't seem to find a motherboard with a non-shared ISA slot and support for a Socket 370 Pentium III 933mhz chip. It seems that boards with seperate/non-shared ISA slots are no longer made, and the ones I've found don't support later chip generations .. can anyone here point me in the right direction?
BilldaCat
You don't need it. The demanding IO is on the PCI bus (North). I don't think the south bridge can feed more than 100MB/s. See http://www.amd.com/products/cpg/server/athlon/chip set.html. It's only IDE (how much can the HDD MECANICS give you) plus USB. No big deal.
Most likely one wouldn't need a WTX case if the board doesn't support AGP Pro 110 (an extra 110W for the video card is more than consumer cards normally worry about)...
--
"It's tough to be bilingual when you get hit in the head."
It would seem that one would prefer the AGP Pro 110W power to not be sucking off of the main power for the CPUs/memory, etc... That explains away a few lines. Easier power routing and regulation for some areas. That would seem to be the greatest separate need.
--
"It's tough to be bilingual when you get hit in the head."
I don't know about the Samurai, but
Micron Mamba chipset (North Bridge only) for the AMD platform is expected to be released in Q3. In addition to DDR SDRAM support, Mamba will also feature 8Mb of L3 cache on the chipset die. The L3 cache will have a sustainable memory bandwidth of 9.6GB/s.
Micron Scimitar chipset for the AMD platform is expected to be released in July. Scimitar is expected to feature a Mamba core with integrated on-die Rendition graphics.
copied word for word from mikeshardware.co.uk (an awesome site for not so publicized chipset/tech news)
You point out why it's always good to root for and support the underdog no matter who it is. If one day AMD becomes dominant then we should help out intel.
Monopolies are bad for the consumer.
War is necrophilia.
Does that Tyan dual PIII have 64-bit PCI?
Does it have *dual* ethernet ports?
Try the Tyan S2510NG: 64-bit PCI, video, dual ethernet, etc: $370
Does the 3950 have dual U160 SCSI? I doubt it for $129. No, it is in fact an U2W card! Compare like with like! Try the Adaptec 39160, at $340!
Of course, the best thing is to compare like with like as much as possible: the tyan S2510U3NG: Dual LAN, U160 (1 channel I think), ATA onboard video, etc: $470. Less than the AMD motherboard, but the AMD motherboard is new, and the processors cost less for more.
Pricewatch sucks. No search within these results option.
The Tyan is a more expensive board, but ~$200 boards are coming using the AMD760MP (or 760MPX) chipset.
The MPX is the same as the MP, but does 64-bit PCI at 66MHz, not 64-bit PCI at 33MHz.
I just want the good stuff from nForce with the good stuff from the 760MPX put together in one great chipset.
I'm not so sure about that. For a high end workstation, you would usually be correct, but this chipset is being touted for servers.
It would not be at all uncommon for a database server to have a couple of the latest SCSI 320 cards running a farm of 15KRPM drives or external RAID chassis. It doesn't take that many of them to saturate a PCI66 I/O channel. Aside from that, you don't want saturated channels on a database server -- you want your I/Os scattered evenly in order to maximize parallelism.
I'm guessing that 760MP boxen will be relatively competitive with some of the 2/4 way systems from Dell, IBM, et. al., but that is relatively low end competition. To really compete as a server, the systems will have to be configured with:
That said, I'm certainly looking into a box for home use. I don't need 5-9 reliability, so I'm just going to be waiting for sane prices...
I do not fail; I succeed at finding out what does not work.
Ace's Hardware posted what I feel is the best review yet of the 760MP. Instead of benching games and business apps, they go for the true workstation-class pieces of software. Caligari TrueSpace, Maya, 3DStudio MAX, SPEC APC, Microstation and Visual C++. They also do a bit of video editing and encoding. But on top of the benchmarks, they also go into detail on the architecture behind the 760MP. They look at how AMD's use of point to point topology is better than Intel's use of a shared bus topology. There's a lot more stuff there, but in regards to space, I'll ley you all find out for yourself.
SYSOP ('sih-sop) n.: the guy laughing at your typing.
ROFL! AMD is a big company just like Intel, and always was. If you think AMD's success started with x86 compatability you're dreaming. A linux analog? Well, maybe if Linux was produced by Oracle.
For god's sake, its MB/sec or MB/s, not MBps! (Unless you mean that the nForce's super cool new bus is slower than ATA/133!)
A deep unwavering belief is a sure sign you're missing something...
While ps is technically correct, the standard usage is MB/sec and Mbps. The main reason most people use it this way is because it is quite easy to mistype MBps as Mbps, and intoduces some ambiguity in the number. While the author may know perfectly well what he means, the reader doesn't know the author personally and thus thinks the author could have made a mistake (as they often do.) Secondly, the ATA-133 my mistake. I meant ATA-100.
A deep unwavering belief is a sure sign you're missing something...
It is at that point I see Alpha systems with dual 256bit busses and get a little jealous...
A deep unwavering belief is a sure sign you're missing something...
I think we should all support AMD products because they show the world that the underdog can suceed. AMD was, only 3 years ago, considered a little player in the CPU market. They had a K-6 CPU that the P-II was killing and Intel triedto squease them (and Cyrix) out of the business with a new slot for CPU's. Intel (like most LARGE companies) didnt think anyone would be upset that an upgrade to there product meant a complete rebuildof the system (MB, RAM, CPU). AMD kept going with the support of a few chipset manufacturers and brought socket 7 all the way to 550 MHz. They won customers that way and when the technology needed it, they had those customers buy there new line of CPU's. AMD is still gaining market share with the Duron and K-7 and I hope that they do just as well in the server market, although I think that will be a tougher fight than the desktop was. A company like AMD gives me hope that Linux (and all GNU software) will be able to someday take the desktop from another HUGE company.
"If ignorance is bliss, why aren't there more happy people in the world?"
What? Using a lower-case 'p' as an abbreviation of "per" is, as far as I know, pretty well recognized as an alternative to the slash. I don't quite get your reference to ATA/133 (whatever that is; perhaps you are thinking of ATA/33?), though...
main(O){10<putchar(4^--O?77-(15&5128 >>4*O):10)&&main(2+O);}
Yeah, me too! I mean, it seems like such an obvious fit--nForce has dual-DDR 4.2 GBps goodness, and an AthlonMP needs 2.1 GBps maximum. 4.2/2.1 = 2! Heh. On second thought, it's not quite so perfect, since some bandwidth for I/O and graphics is good to have. Um, Nvidia, could you add another 2.1 GBps to that? Thank you. ;^)
main(O){10<putchar(4^--O?77-(15&5128 >>4*O):10)&&main(2+O);}
After spending yesterday reading about all the good stuff Nvidia has crammed into the nForce, including the nice 800 MBps "HyperTransport" link between their versions of the north and south bridges, I threw myself over these long-awaited 760MP exposes, to see what AMD use. I'm more than a little surprised (and disappointed) to find that they went with the "good-old" PCI interconnect, limited to a measly 266 MBps (if it's 64-bit). The weirdness increases when you realise that Nvidia didn't actually develop HyperTransport themselves--it's licensed from (wait for it) AMD!
I guess the reason is that HyperTransport is too recent a development for AMD to include it in the 760MP, which has been under development and testing for like two years, but still... It's a shame. It seems that even the upcoming "mainstream" SMP chipset, the 760MPX, won't include HyperTransport.
main(O){10<putchar(4^--O?77-(15&5128 >>4*O):10)&&main(2+O);}
You have to factor in the CPU prices.
From PriceWatch:
TYAN DUAL AMD $565
(includes onboard video, lan, scsi)
Duron 900Mhz $64 X 2
Total $689
TYAN DUAL PIII $222
(includes onboard video, lan)
PIII 1Ghz $184 X 2
Adaptec 3950 SCSI $129
Total $719
In early benchmarks, the Duron 900mhz is comparable to the 1Ghz PIII.
Tyan is only painful if you have zero use for SCSI. (everyone needs LAN, and onboard video is just an cheap annoyance).
(of course the flaw in this argument is RAM prices, but if you buy namebrand stuff, it turns out the registered isn't so much more)
So when do I get one of these sweet boards with built in hardware IDE RAID?
Anyone else interested?
I don't have $$$ for scsi.
Spoon not. Fork, or fork not. There is no spoon.
This is great, first AMD attacked Intel w/ a rival to it's PIII, the Athlon, then w/ the Duron vs. Celeron, which gave them major exposure in the desktop pc market, and now, w/ dual processor support, they are attacking the "big money" market, that of servers. I'm sure I speak for many here when i say, 'way to go AMD!'. They really deserve the credit. I for one will be switching our servers from dual PIII servers to dual athlons.
E.
-
-
This Post has been brought to you by the letter "E".
Uh, it's not an ATX motherboard, it's WTX, which has different standards.
:-)
Don't worry, I'm sure once the 760MPX is released we'll see some very consumer oriented motherboards hit the market. MSI has already announced that they want to do a sub-$200 motherboard... Wait for that one!
well the last time I checked the pinout on my powersupply it didn't include 110 volt wiresso I would imagine that is what the extra 4 wires are for
Jon
This is awesome. But it looks like is has special requirements for power supplies and chassis. Does anyone know of any compatible cases (preferably rack-mount with SCA hot swap bays) and power supplies for this baby?
The review including the benchmarks on Tom's Hardware are detailed, of course, but what we really want to know is:
what if we benchmarked a Beo... well, never mind.
Did you catch the thing about non-standard pinouts on the power connectors?
BLAST IT!
As the owner of two Tyan (mid Rev4 Tomcat I, later Trinity 1590S) boards, they really $%^& this one up. Non-standard PS2 mouse connectors, non-standard serial port connectors, non-standard USB connectors.
With these connectors it's not too bad, because the Tyan-pinout ones aren't much more, or it's not difficult to modify a standard one. But to mess up on the power supply connector...
I still haven't been able to get DMA running on the 1590S, on either stock Redhat kernels or using the Jumbo IDE patch. At some level, others have their MVP3's running DMA.
I have been pleased with the stability of Tyan boards, but between connector issues and the DMA troubles I've been having, it no longer feels safe as a 'default' decision.
The living have better things to do than to continue hating the dead.
---=-=-=-=-=-=---
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
... time to break those piggybanks
I was just reading yesterday's article on Crush and suddenly all the hardware sites slowed to a crawl
Hrrmph
Special Relativity: The person in the other queue thinks yours is moving faster.
Why haven't anyone tested the motherboard (with all its features) under Linux?
They have a cursory look at Linux on page 17 of the Anand article. Not as much info as you'd like, but enough to reckon they have it working on the Mobo with little problem
Special Relativity: The person in the other queue thinks yours is moving faster.
rr
Quidquid latine dictum sit, altum videtur.
Take a look at the numbers on this graph at Anandtech: http://images.anandtech.com/reviews/chipsets/amd/7 60MP/linux.GIF (copy and paste the link yourselves) that the 2P configurations compiled the kernel even only using 1 thread at about +10% speed of the 1P setups. Now look at the performance of the single threaded 1P version versus the 2P two thread versions. There is a 101% increase in performance. I think we really ought to recognize the guys who contribute to the kernel that have written such efficient multithreading a little more.
I don't understand why this is never mentioned in the reviews. How did support for MP Athlon just appear? From my understanding, AMD is using the EV6 multiprocessor protocol which is incompatible Intel's patented APIC protocol. So how does Windows 2000 just work?
- Database Server Performance: 17.6% Faster
- 3D Rendering Performance: 14% Faster
- Image Editting Performance: 6.1% Faster
- Workstation Performance (Overall): 22.1% Faster*
- Linux Performance (Total): 12% Faster
- IT/Constant Computing Performance (Average): 17% Faster
- Overall System Performance: 8.6% Slower**
* - The Xeons did outperform the Athlons on the Photoshop 4 portion of the workstation performace scores by 11.4%.** - The Overall System Performance numbers ended up that way due to the Xeons' 20% advantage over the Athlons on the Internet Content Creation benchmarks and the basically even performance on the Office Productivity benchmarks.
Portable versions of Firefox, GIMP, LibreOffice, etc
even though it only operates at 200 mhz as opposed to 266 mhz, it rocks. I highly recommend the orbital cooling fan, though I had to modify it to get it to fit on the motherboard. I also recommend dual case fans, and air conditioning. even on a cold day it heats up my room. the other flaw with the gigabyte board is that it won't load windows 2000, and gigabyte has zilch for tech support.
CPU power has ramped up considerably since the Athlon debued and gave Intel a scare.
mhz has ramped up considerably; the jury is still out on cpu power
the animal doesnt even have opposable thumbs, focker!
...did AMD/Tyan go for yet another PSU connector system? The P4 uses the normal ATX connector plus a second 4 pin block for more juice. Surely it'd be more economical to use this system as many PSUs have this connector already.
.. ?
BTW, is it me, or is the main connector location on the tyan board really bad
No Norm, those are your safety glasses; I'll wear my own thanks...
This is truly a very good example of why competition is important. Granted, we've seen drastic improvements in processors (even before AMD applied pressure), but most of that was just in clock speed. Now we're seeing some incredible advances in processor design beyond just the clock speed race. And consumers are starting to become aware that there is much more to be told about its processor than its clock speed.
Way to go, AMD. Keep turning the screws. Things can only get better for the consumers.
GreyPoopon
--
GreyPoopon
--
Why is it I can write insightful comments but can't come up with a clever signature?
test
Gotta hand it to em over at Anandtech ... they are doing the right thing. Either that, or the performance of the dual Athlons so overpowers whatever Intel can deliver -- or both :=)
------ Give a man a flame, keep him warm for a night. Set a man on fire, heat him up for life.
I really miss my old dual pentium 100MHz. SMP really rocks, even when used in a workstation. Everything is "snappier" because other processes does not bug down a new one.
But the price for a dual athlon 1GHz... I think I will wait until the price is reasonable for a dual "something" ~2GHz
Why haven't anyone tested the motherboard (with all its features) under Linux? Does the SCSI work? What about the onboard sound stuff? Is the asic supported by which kernels? etc.