Dothan and Turion Compared and Explained

All about them.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Dothan and Turion Compared and Explained

Post by ~El~Jefe~ » Tue Aug 30, 2005 7:42 am

Really nice article. I would go read it fully.

The end benchmarks lack a tiny bit in fairness as the Dothan is using 100mhz bus vs 133 mhz bus like it would in optimal situations, or in a desktop environment. Besides that, it is a real eye opener for how strong the lil 754 Turion.

http://anandtech.com/news/shownews.aspx?i=24811

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Tue Aug 30, 2005 7:47 am

Thanks, ill read it later. Rihgt now im so incredibly tired and mbored i dont want to read anything.

Oh, you asked in another thread but ill answer here. No i havent found a car yet, might also stick with the current one and research the possibility of adding a supercharger. Though i want RWD, but insurance doesnt cover track driving so thats a bummer. so maybe ill keep this one, and get a cheap rwd car and use that to do track days. or just get an Lincoln Continental -63 convertible or something an cruise with it.

I really dont care at the moment.

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Tue Aug 30, 2005 11:22 am

Finally, a review that does the Turion some justice! GamePC's comparison was fundamentally flawed and promoted many misconceptions about Turion power consumption and performance.

Everything seems pretty spot-on in the Laptop Logic review. The only thing that really caught me by surprise was the Batterymark "life" benchmark. Don't know exactly why the Turion was able to pull ahead in this situation. Perhaps PowerNow! is more efficient than SpeedStep? Or maybe lowering frequency/voltage does not have as significant an effect on Dothan's power consumption, as it already uses very little power? I believe I remember Ralf pointing out something to this extent based on his experience with Pentium-M desktop platforms.

cotdt
Posts: 295
Joined: Tue May 10, 2005 3:30 pm

Post by cotdt » Tue Aug 30, 2005 11:39 am

Load is what matters from heat production perspective, not idle. At SPCR most of us are interested in how suitable these are in silent desktop systems.

So the Turion still doesn't even come close to the Dothan. 1:51 versus 2:32 battery life is a very significant difference once you factor in the power consumption from the other components. It translates to using 30-40% more power, and thereby more heat, than the Dothan. Also, this isn't the best chipset for the Dothan, the other chipsets perform far better.

So I don't see anything to rave about. AMD still has a lot of catching up to do. I still have the feeling that undervolting and underclocking my Venice can produce the same results as that Turion. That said, Turion does support 64-bit.

Still, what matters at the end of the day is that you can run one CPU passive (true passive) w/Ninja heatsink while the other you cannot. Turion has yet to pass the treshold for passive cooling.

qviri
Posts: 2465
Joined: Tue May 24, 2005 8:22 pm
Location: Berlin
Contact:

Post by qviri » Tue Aug 30, 2005 11:47 am

frostedflakes right above you is running his (slightly undervolted) Turion MT-34 with a fanless Scythe NCU-2005. His S12 is the only fan in that system.

cotdt
Posts: 295
Joined: Tue May 10, 2005 3:30 pm

Post by cotdt » Tue Aug 30, 2005 11:56 am

qviri wrote:frostedflakes right above you is running his (slightly undervolted) Turion MT-34 with a fanless Scythe NCU-2005. His S12 is the only fan in that system.
Ahhh interesting... but it's undervolted to 1.075. I don't know if all the Turions could do that. Also, how would it work with an Antec Phantom 350? To be truely passive the whole system needs to be passive.

qviri
Posts: 2465
Joined: Tue May 24, 2005 8:22 pm
Location: Berlin
Contact:

Post by qviri » Tue Aug 30, 2005 12:02 pm

I daresay "truly passive" is impossible in a standard ATX case anyway.

Certainly very impractical and not worth the effort.

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Tue Aug 30, 2005 12:12 pm

qviri does not lie. The processor stays at a cool 39*C while folding.

Also, idle/near-idle power consumption is very important. Most people use laptops for business applications (word processing, internet, etc.), which only slightly load the processor. Sure, the Dothan takes the lead at maximum load, but also keep in mind that the AMD notebook is using a 35w Turion. I'd expect a 25w to be much more competitive with the Dothan in this area.

And which chipset would you prefer? As far as I know, there are only two Dothan chipsets: 855Gme and 915Gm. The Centrino laptop tested uses the most recent 915Gm chipset, which supports PCI-E and DDR2.

Don't get me wrong, though. Overall, I have always considered Centrino to be the better platform. But I was getting sick of people basing Turion's worth on GamePC's flawed review. As long as people accept Turion for what it is, I will be happy. Turion is NOT a Pentium-M killer. Just another option for mobile users that has its own set of strong points and weak points. ;)

P.S. And about the "truly passive" PC, I intend to try that out sometime soon. I plan to replace my S12 with an internat DC-DC ATX converter. Becuase of the high efficiency of the DC-DC conversion, the internal part of the power supply shouldn't give off very much heat. The external AC/DC conversion is the inefficient part, but this won't be a concern, as the brick will be outside the case releasing its heat to the ambient air.

I doubt I'd be able to run completely passive during load, but it may be possible at idle. I will have a 120mm exhaust in the back that will turn on when necessary. But even then, it shouldn't have to spin very fast to dissipate the heat from my PC. The Yate Loon in my S12 is only ramps up to ~785RPM while gaming, and it has to cool the power supply as well.

I know it's not practical. My PC is already silent from where I'm sitting. I just like to tinker with PCs and try new/different things. :)

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Re: Dothan and Turion Compared and Explained

Post by rpsgc » Tue Aug 30, 2005 1:47 pm

~El~Jefe~ wrote:Really nice article. I would go read it fully.

The end benchmarks lack a tiny bit in fairness as the Dothan is using 100mhz bus vs 133 mhz bus like it would in optimal situations, or in a desktop environment.
In the article's comments.
fbrdphreak wrote:In the part where we compare the Turion Compaq V2000Z to the PM Compaq V2000, yes the V2000 is using the 400MHz FSB 855 platform. An article will be launching soon detailing the benefits of going to the Sonoma (533MHz FSB) platform, and I can tell you that it is not a huge performance increase by any means. So using the 400MHz FSB platform was sufficient for the overall POINT of that comparison, that is a price comparison. The two models were so similar it was the best way to compare value in a mainstream notebook.

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Tue Aug 30, 2005 3:39 pm

cotdt wrote:So the Turion still doesn't even come close to the Dothan. 1:51 versus 2:32 battery life is a very significant difference once you factor in the power consumption from the other components. It translates to using 30-40% more power, and thereby more heat, than the Dothan.
You can also tell from the review that PowerNow is more efficient than SpeedStep at idle.
In an ironic twist of fate, the Turion CPU defeats Dothan at its own game. Since the "life" benchmark tests battery life at a mostly idle CPU state, maximum power consumption does not matter. With a 10.5% lead in the life test, the Turion CPU and AMD’s PowerNow! Technology provide a more efficient platform when you consider real world usage. We will delve into the details of what platform is better under what circumstances, but the tests above clearly show the Pentium M providing better high-load power consumption and Turion offering longer battery life during more realistic non-intensive usage.

Likif
Friend of SPCR
Posts: 189
Joined: Tue Nov 26, 2002 11:40 am
Location: Copenhagen, Denmark

Post by Likif » Wed Aug 31, 2005 1:07 am

Here in Denmark Turion laptops are slightly more expensive than Dothan Centrinos.

Now why, now why is that? mmm?

I'd like to see the Sempron-M compared to the old Centrino.

3waybar
Posts: 5
Joined: Wed Aug 31, 2005 1:45 am
Location: Sw Southpole
Contact:

Post by 3waybar » Wed Aug 31, 2005 1:54 am

Why doesnt the reviewer put equal effort of replacing the Dothan's RAM
preferably to some DDR2 with 3-2-2 latency ?

http://pcmoddingmy.com/content.php?review.cat.41

lots of reviews on ex. Patriot, OCZ, Mushkin DDR2

they state memory performance increases of about 4-10 %
enough for the Dothan to close the gap in all memory tests :roll:

Bar81
Posts: 261
Joined: Thu Sep 04, 2003 4:19 pm
Location: Dubai

Post by Bar81 » Wed Aug 31, 2005 10:29 am

The article has no application whatsoever to the desktop arena. Put a 533 FSB Dothan on a P4C800-E Deluxe, up the FSB to 200+ with PAT enabled, drop in some CAS2 RAM and watch it pound the turion mercilessly in gaming and office and win almost all the rest. I'm not quite sure where you guys are dreaming up that the Turion is power competitive with the P-M, there's a full 30% delta between the two chips *when you're actually using them*. Not that the Turion can even come close to the overclocking potential of the P-M. The bottom line is the P-M is by far the superior choice for a quiet system over the Turion for those that want power and silence.

Anyway, both of these chips should be no match for the beautiful Yonahs coming out in dual core flavor and die shrunk with all the new enhancements. Best news is that they are S479 so theoretically those with S479 setups should be able to run them with a simple BIOS update.

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Wed Aug 31, 2005 10:33 am

Go go fanboism :roll: You do know there's a 25W Turion don't you? And that Intel's 27W TDP are no less than AMD's 35W TDP? Pff.... desktop? We're talking about MOBILE here. For desktop there's the Athlon 64, yeah?

Just wait until there's a chipset fully optimized for the Turion, then we'll talk.
People underestimate the Turion wayyyy too much :roll:

Bar81
Posts: 261
Joined: Thu Sep 04, 2003 4:19 pm
Location: Dubai

Post by Bar81 » Wed Aug 31, 2005 11:29 am

rpsgc wrote:Go go fanboism :roll: You do know there's a 25W Turion don't you? And that Intel's 27W TDP are no less than AMD's 35W TDP? Pff.... desktop? We're talking about MOBILE here. For desktop there's the Athlon 64, yeah?

Just wait until there's a chipset fully optimized for the Turion, then we'll talk.
People underestimate the Turion wayyyy too much :roll:
Yeah, glad you're intelligent enough to equate the truth with fanboism. Buy a clue.

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Wed Aug 31, 2005 11:40 am

Bar81 wrote:Yeah, glad you're intelligent enough to equate the truth with fanboism. Buy a clue.
The thruth? Please do READ the article, and then talk about "the thruth". Saying Turions sucks, even though this article plainly states otherwise is not fanboism, riiight. Whatever floats your boat.

I'm intelligent enough to know that Turions are competitive and that they're just first gen and do not have a highly optimized chipset, unlike Dothan.

cotdt
Posts: 295
Joined: Tue May 10, 2005 3:30 pm

Post by cotdt » Wed Aug 31, 2005 11:49 am

Bar81 has hit the nail. =)

The 25W Turions are just selected 35W Turions with the voltage lowered to 1.2V. If you're undervolting, it hardly makes any difference. Let's keep the discussion on desktops, shall we? Most at SPCR are interested in desktops. What is important is heating, not power consumption. Here Turion doesn't stand a chance compared to the Dothan.

IsaacKuo
Posts: 1705
Joined: Fri Jan 23, 2004 7:50 am
Location: Baton Rouge, Louisiana

Post by IsaacKuo » Wed Aug 31, 2005 12:00 pm

cotdt wrote:Still, what matters at the end of the day is that you can run one CPU passive (true passive) w/Ninja heatsink while the other you cannot. Turion has yet to pass the treshold for passive cooling.
Have I missed something? This seems to imply that you can run a Pentium M with a Ninja heatsink. On what motherboard? The only one I'm familiar with is the one reviewed here on SPCR, which has a proprietary heat sink mount.

I don't follow computer hardware news very much, so I'm honestly curious.

cotdt
Posts: 295
Joined: Tue May 10, 2005 3:30 pm

Post by cotdt » Wed Aug 31, 2005 12:20 pm

IsaacKuo wrote:
cotdt wrote:Still, what matters at the end of the day is that you can run one CPU passive (true passive) w/Ninja heatsink while the other you cannot. Turion has yet to pass the treshold for passive cooling.
Have I missed something? This seems to imply that you can run a Pentium M with a Ninja heatsink. On what motherboard? The only one I'm familiar with is the one reviewed here on SPCR, which has a proprietary heat sink mount.

I don't follow computer hardware news very much, so I'm honestly curious.
Any motherboard, you just have to do some custom mounting. Pretty easy for DIYers. I tried it with open case but no airflow and temps stayed pretty low. Too bad the Dothan system belonged to a friend and not me. There are reports that undervolted Turions also work, but there are always some other fan in the system so I don't know if they could really be run passive. Dothans can run hot w/o crashing unlike Desktop CPUs, they can handle 100C. This makes them a great candidate for passive cooling.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Wed Aug 31, 2005 3:19 pm

eh. the article clearly stated that the turion was faster in just about every way.

or at least equal, but it is cheaper.

The reason I said that the 100 vs 133 bus comparison was needed is that the % difference between the two processors was kinda small, yet this would make the sonoma a little faster, which would mean that it could be actually faster or as fast as a turion.


I do recall that cool n quiet is more efficient in real life usage of a chip wattage saving wise than speedstep. I forget why though, the topic hasnt been talked about much in a year.

As an aside: Why isnt the top dual core pentium using speedstep? It is the only feature the "HT" model doesnt have on it.

mathias
Posts: 2057
Joined: Sun Jul 11, 2004 3:58 pm
Location: Toronto
Contact:

Post by mathias » Wed Aug 31, 2005 4:20 pm

cotdt wrote:Load is what matters from heat production perspective, not idle.
It matters if you use a thermally controlled fan. A power supply that ramps up somewhat under load probably impllements this quite often without even being noticed.
cotdt wrote:So I don't see anything to rave about. AMD still has a lot of catching up to do.
How demanding. AMD caught up plenty of times, they didn't really benefit much from it.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Wed Aug 31, 2005 9:02 pm

Socket M2 hasnt even come out yet. No one has really tested a 65 nm chip besides shows, and we all know how accurate and unbiased those relults are. :roll:

turion is also still single channel. Dothan/sonoma didnt benefit from it much, but turion certainly would.

I just wish i could use a 25watt turion in this machine here. that'd be nice. id oc .2+ghz and bet it would still be the coldest thing on load.

Anyone have luck with 90nm turions in 754 boards?

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Wed Aug 31, 2005 9:31 pm

Vandi423 on XtremeSystems.org recently picked up an MT-37. It's running at 2.0GHz 1.07V.

Then there's my Turion, which I'm sure you've already read about on the forums. It's a very early Turion (part of the first batch), and doesn't underclock as well.

Mikael
Posts: 206
Joined: Mon Dec 06, 2004 3:12 am
Location: Gothenburg, Sweden

Post by Mikael » Thu Sep 01, 2005 2:57 am

cotdt wrote:What is important is heating, not power consumption. Here Turion doesn't stand a chance compared to the Dothan.
Huh!? As far as I know, pretty much all the power that goes into a chip is turned into heat, so that statement is a little off.
Bar81 wrote:The article has no application whatsoever to the desktop arena. Put a 533 FSB Dothan on a P4C800-E Deluxe, up the FSB to 200+ with PAT enabled, drop in some CAS2 RAM and watch it pound the turion mercilessly in gaming and office and win almost all the rest.
Yeah, and give a the Turion dual channel memory, a desktop chipset and CL2 memory and you'll see it pretty close to the Dothan. Well, that's unless you like to read GamePC's articles...
Bar81 wrote:I'm not quite sure where you guys are dreaming up that the Turion is power competitive with the P-M, there's a full 30% delta between the two chips *when you're actually using them*. Not that the Turion can even come close to the overclocking potential of the P-M. The bottom line is the P-M is by far the superior choice for a quiet system over the Turion for those that want power and silence.
The real comparison would be Dothan versus Turion MT (25W). Dothan is still likely to be more efficient, but not by a huge margin. It's also hard to draw any definitive conclusions without knowing the power consumption of the rest of the system.
3waybar wrote:Why doesnt the reviewer put equal effort of replacing the Dothan's RAM
preferably to some DDR2 with 3-2-2 latency ?
Isn't that very obvious? The only thing they do with the Turion is to replace the RAM with standard PC3200 SODIMMs that should be fitted in all Turion systems. Putting some really expensive low latecy DDR2 in the Dothan system would hardly make things fair. Either way, I'm quite sure that those high performance modules your talking about aren't available in SODIMM flavor.

Mats
Posts: 3044
Joined: Fri Dec 26, 2003 6:54 am
Location: Sweden

Post by Mats » Thu Sep 01, 2005 2:47 pm

It's a very good review but I think it's most interesting for mobile users. I mean, how many here are actually considering T64 for desktop unless you already have a mobo that you know will work with it?
A64 should be enough for everyone (who wants AMD)!
ddrueding1 wrote:Here's my experience underclocking my Venice-core CPU.

This example is a 3500+, whose maximum multiplier is 11 and FSB is 200Mhz (therefore 2.2Ghz CPU)

I've been playing with CrystalCPUID to see what the processor can do and have made some really neat discoveries. First thing I did was OC the FSB to 228 to give a max clock of 2.5Ghz+.

1. At it's minimum voltage of 0.8v it produces virtually no heat at it's minimum multiplier (4x == 900Mhz+)
2. Still at it's minimum voltage (0.8v) it runs Prime95 100% stable at a 9x multiplier (2Ghz+)
3. It can run at it's stock clock speed (2.2Ghz) at 1v (default is 1.4v)

The thing that really blew me away was the 2Ghz+ clock speed running at only 0.8v. I'm not sure how much power this thing is actually drawing, but it is doing quite well completely passively! I'm quite thrilled.
This is outside the case, but I'm very impressed anyway! 0.8 V???

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Thu Sep 01, 2005 3:33 pm

I agree, Turion just isn't worth it for most people when a 3000+ OEM can be had for ~$120 and will undervolt like mad. A Turion with comparable performance would probably set you back close to $200.

dragmor
Posts: 301
Joined: Sun Jul 10, 2005 7:54 pm
Location: Oz

Post by dragmor » Thu Sep 01, 2005 3:41 pm

Mats wrote:It's a very good review but I think it's most interesting for mobile users. I mean, how many here are actually considering T64 for desktop unless you already have a mobo that you know will work with it?
A64 should be enough for everyone (who wants AMD)!
ddrueding1 wrote:Here's my experience underclocking my Venice-core CPU.

This example is a 3500+, whose maximum multiplier is 11 and FSB is 200Mhz (therefore 2.2Ghz CPU)

I've been playing with CrystalCPUID to see what the processor can do and have made some really neat discoveries. First thing I did was OC the FSB to 228 to give a max clock of 2.5Ghz+.

1. At it's minimum voltage of 0.8v it produces virtually no heat at it's minimum multiplier (4x == 900Mhz+)
2. Still at it's minimum voltage (0.8v) it runs Prime95 100% stable at a 9x multiplier (2Ghz+)
3. It can run at it's stock clock speed (2.2Ghz) at 1v (default is 1.4v)

The thing that really blew me away was the 2Ghz+ clock speed running at only 0.8v. I'm not sure how much power this thing is actually drawing, but it is doing quite well completely passively! I'm quite thrilled.
This is outside the case, but I'm very impressed anyway! 0.8 V???
Ive got a A64 3500+ in a SN95G5 (which tends to undervolt). I havent finished testing, but my results so far are:

2200mhz @ 1.2v (default 1.4v)
2000mhz @ 1.025v
800mhz @ 0.8v

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Fri Sep 02, 2005 3:42 pm

any via 800 pro chipset board use a turion??

now that would be my upgrade.

754 abit kv8 Pro is what i have.... shrugs......

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Fri Sep 02, 2005 9:13 pm

I have an ASUS K8V-X on the way for my Turion. It uses Via K8T800.

Turion should work in any board with a rev. E compatible Sempron BIOS, though. The Turion cores and rev. E Sempron cores aren't identical, but they're similar enough that the board should still run fine with a Turion. Worse case scenario you'd probably have to manually set voltage, multiplier, etc. in the BIOS.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Sat Sep 03, 2005 7:07 am

frostedflakes wrote:I have an ASUS K8V-X on the way for my Turion. It uses Via K8T800.

Turion should work in any board with a rev. E compatible Sempron BIOS, though. The Turion cores and rev. E Sempron cores aren't identical, but they're similar enough that the board should still run fine with a Turion. Worse case scenario you'd probably have to manually set voltage, multiplier, etc. in the BIOS.
doesnt the chip itself automatically tell the board what to set itself at?

yay or nay?

I have an Abit kv8 pro. if i were to get a new board, i then wouldnt bother with turion. I would just get a venice core. so.... i wonder if it will work?

the board isnt undervoltable in bios.

Post Reply