Athlon 64 3000+ to 4200+ X2 with OC?

[21stHermit]

I have a single core Athlon 64 socket 939, which I'm thinking of upgrading to this AMD Athlon 64 X2 4200+ Manchester Then I'd like to overclock it if its worth the effort. I'll only use the computer to compile text map files into binary, no gaming, no daily use. I have ~1000 hours of compiling in the next six months.

My Questions?
How much can I expect to OC the X2 4200+ Manchester?
Will I have to re-register XP with the CPU swap?
Other alternatives/suggestions?

Some Background on my current system and benchmarks:

Current System/Components:
CPU: Athlon 64 3000+ Winchester, 1.8GHz
MB: MSI K8T Neo2-F
RAM: 2x1GB Patriot DDR400
VGA: Generic AGP card
HSF: Have a Big Typhoon NIB

Here's a Tom's CPU Chart comparing my current CPU to the closest match I could find Toledo, not Manchester. So if WinRAR is an equivalent benchmark, then I should see a ~40% reduction in compile time w/o OC.

Here's a recent benchmark I ran comparing my PC's using the text to binary compiling software:

My PC's:
1. Dell E1505 w/ Core Duo T2300 1.66GHz, 2GB DDR2
2. Celeron D 360 Cedar Mill, 3.46GHz, 2GB DDR2
3. Athlon 64 3000+ Winchester, 1.8GHz, 2GB DDR

Benchmark Test:
Compile 60.5MB of text files to 15 finished binary files totaling 10.4MB, this is a 1 x 1 degree extent of 40-foot contour maps.
E1505........ 52 min
CeleD360... 30 min
Athlon64.... 19 min
A 64 X2 .... 18 min Bummer

Thanks
Hermit

Edit: Added X2 4200+ Manchester time.

 

[MarcVenice]

I'm not sure if unzipping something with winrar is anything like what you do, but when I do that, my cores don't run at 100%, more like 30-40%. I figured the HD didn't allow for faster unzipping or whatever. It would be a worthwhile upgrade in my opion though, for 60$ you can't really go wrong with it ...

 

[21stHermit]

Originally posted by: MarcVenice
I'm not sure in unzipping something with winrar is anything like what you do, but when I do that, my cores don't run at 100%, more like 30-40%. I figured the HD didn't allow for fasting unzipping or whatever. It would be a worthwhile upgrade in my opion though, for 60$ you can't really go wrong with it ...

Thanks Marc,

I assumed that was zipping, not unzipping, but not an expert. In my case I'll be running off a 10/100 NAS, the files are very small relative to the CPU time so the HD is a minor issue.

Since posting and searching, that CPU was selling for ~$300 a year ago. Oh what C2D has done to the market!!!

Hermit

 

[hans007]

i am doing this same upgarde for my brother, hopefully itll keep his computer useable for another year.

i think newegg had it for $60 shipped. i would assume it'd be a lot faster even on single threads the x2 is about 20% faster than a 3000

 

[FarENheiT]

How about an even SILLIER question?

I have an Athlon 64 4000+ (socket 939), and was looking to max out my motherboard. It takes up to an Athlon 64 X2 4800+ Toledo (), however I can't find those on the Internet for less than $300. So, for $70, is it worth it to go from a 4000+ to a 4200+ x2? System would be used for some gaming, design, video, etc.

Current Rig:
Gigabyte K8N51PVMT-9 motherboard
Athlon 64 4000+ (socket 939)
2GB RAM
ATI All-In-Wonder X1800

Thanks!

 

[myocardia]

Hermit, since you already know you'll be overclocking, do yourself a favor, and spend the extra $10 for the overclocker of the two, the 4200 Toledo. Most of the Manchesters topped out between 2.4 & 2.5 Ghz. It's only the Toledo's that are doing 2.7 Ghz & faster.

 

[VirtualLarry]

Originally posted by: myocardia
Hermit, since you already know you'll be overclocking, do yourself a favor, and spend the extra $10 for the overclocker of the two, the 4200 Toledo. Most of the Manchesters topped out between 2.4 & 2.5 Ghz. It's only the Toledo's that are doing 2.7 Ghz & faster.

$10? If only!

Manchester $60
Toledo $115

 

[MarcVenice]

Just buy this CPU: http://www.newegg.com/Product/...x?Item=N82E16819103052

It's pretty much a gauranteed 2.6ghz overclock, and if your mobo can do it, it can probably do higher as well. Got one myself, so I know what I'm talking about.

@Farenheit, do the same as mentioned above, simply overclock it. Otherwise you'll be seeing a loss, even though it would be a minor one, in most single threaded apps.

 

[myocardia]

Originally posted by: VirtualLarry

Originally posted by: myocardia
Hermit, since you already know you'll be overclocking, do yourself a favor, and spend the extra $10 for the overclocker of the two, the 4200 Toledo. Most of the Manchesters topped out between 2.4 & 2.5 Ghz. It's only the Toledo's that are doing 2.7 Ghz & faster.

$10? If only!

Manchester $60
Toledo $115

Well, I looked it up before I typed that, and there was only a $10 difference. They're obviously running low on them now. But, I'd rather have a 3800 Toledo over a 4200 Manchester.

 

[WildHorse]

the closest match I could find Toledo, not Manchester.

As I understand it, the 4200+ Toledo (which I have) contains 2 cores which each have a Manchester core.

Toldeo is just a name they gave the chip with 2 Manchester cores. They aren't different things.

 

[Gillbot]

Originally posted by: myocardia
Hermit, since you already know you'll be overclocking, do yourself a favor, and spend the extra $10 for the overclocker of the two, the 4200 Toledo. Most of the Manchesters topped out between 2.4 & 2.5 Ghz. It's only the Toledo's that are doing 2.7 Ghz & faster.

Who says the manchesters can't hit 2.7G?

AMD A64 x2 4200+ OEM Hit 2740 on AIR without trying (YMMV) with THIS Arctic Cooling Freezer 64 Pro

I didn't even try hard to get a good OC. Just a quick change on the HT and upped the FSB to 250 and it went right in no problems. Then I ran it through superpi and supcom marks to check for stability.

 

[Nathelion]

How about the opty 165? At rated speeds it doesn't quite measure up, but they OC better. (I am currently running one at 2.8 with a voltage bump, so I'm obviously a bit biased).

 

[myocardia]

Originally posted by: scott
As I understand it, the 4200+ Toledo (which I have) contains 2 cores which each have a Manchester core.

The 4200 Toledo has Toledo cores; the 4200 Manchester has Manchester cores. The only reason they have a name after their number, since the number is their actual name, is to denote which cores they contain.

Toldeo is just a name they gave the chip with 2 Manchester cores. They aren't different things.

They're different processors, produced at different times, and have different steppings.

Originally posted by: Gillbot
Who says the manchesters can't hit 2.7G?

With overclocking, as always, your mileage may vary. But, if you look around, and find 10 people running 2.7-2.8 Ghz, with an X2 4200, the majority of them will own Toledo's, not Manchester's. I suppose I shouldn't have said they top out around ~2.5 Ghz, though quite a few of them do/have, at least from what I've seen.

Then I ran it through superpi and supcom marks to check for stability.

That's not how you check for stability, with an overclock. My Opteron 170 will run SuperPi @ 3.0 Ghz, with less vcore than it requires to be stable @ 2.8 Ghz.

 

[Gillbot]

Originally posted by: myocardiaThat's not how you check for stability, with an overclock. My Opteron 170 will run SuperPi @ 3.0 Ghz, with less vcore than it requires to be stable @ 2.8 Ghz.

It ran through orthos for 12 hours as well so that's stable enough for me. I went through quite a few supcom games as well. If all I use this PC for is email and supcom, to me that's stable.

I was just suprised as I didn't really even try to push this cpu at all. I just set it to 250fsb and let it rip. No vcore, chipset or ram voltage changes. They were all left to auto and cpuz reported cpu voltage at 1.312v.

 

[soonerproud]

Originally posted by: 21stHermit
I have a single core Athlon 64 socket 939, which I'm thinking of upgrading to this AMD Athlon 64 X2 4200+ Manchester Then I'd like to overclock it if its worth the effort. I'll only use the computer to compile text map files into binary, no gaming, no daily use. I have ~1000 hours of compiling in the next six months.

My Questions?
How much can I expect to OC the X2 4200+ Manchester?
Will I have to re-register XP with the CPU swap?
Other alternatives/suggestions?

Some Background on my current system and benchmarks:

Current System/Components:
CPU: Athlon 64 3000+ Winchester, 1.8GHz
MB: MSI K8T Neo2-F
RAM: 2x1GB Patriot DDR400
VGA: Generic AGP card
HSF: Have a Big Typhoon NIB

Here's a Tom's CPU Chart comparing my current CPU to the closest match I could find Toledo, not Manchester. So if WinRAR is an equivalent benchmark, then I should see a ~40% reduction in compile time w/o OC.

Here's a recent benchmark I ran comparing my PC's using the text to binary compiling software:

My PC's:
1. Dell E1505 w/ Core Duo T2300 1.66GHz, 2GB DDR2
2. Celeron D 360 Cedar Mill, 3.46GHz, 2GB DDR2
3. Athlon 64 3000+ Winchester, 1.8GHz, 2GB DDR

Benchmark Test:
Compile 60.5MB of text files to 15 finished binary files totaling 10.4MB, this is a 1 x 1 degree extent of 40-foot contour maps.
E1505........ 52 min
CeleD360... 30 min
Athlon64.... 19 min

Thanks
Hermit

I made a similar upgrade earlier this year from a Athlon 64 3200 to a X2 4400. For the price, the X2 is a very worthy upgrade even if you don't overclock the CPU. For what you will be using it for there will be a really good gain on performance.

The main difference between the 4200 and the 4400 is I have an extra Meg of cache. There is not much of a performance difference between the two in benchmarks.

 

[21stHermit]

Originally posted by: soonerproud
I made a similar upgrade earlier this year from a Athlon 64 3200 to a X2 4400. For the price, the X2 is a very worthy upgrade even if you don't overclock the CPU. For what you will be using it for there will be a really good gain on performance.

The main difference between the 4200 and the 4400 is I have an extra Meg of cache. There is not much of a performance difference between the two in benchmarks.

I wish you were right, I just did a bench today on my most important application, its even multi-threaded for dual cores and I got an insignificant 1 min improvement, ~5%.

So far the Athlon 64 X2 4200+ Manchester is a big disappointment.

BTW, in POST it did show up as a Athlon 64 X2 4200+ so I didn't get the wrong chip.

Thanks everyone, for all the posts.

Hermit

 

[betasub]

Originally posted by: 21stHermit
I just did a bench today on my most important application, its even multi-threaded for dual cores and I got an insignificant 1 min improvement, ~5%.

So far the Athlon 64 X2 4200+ Manchester is a big disappointment.

BTW, in POST it did show up as a Athlon 64 X2 4200+ so I didn't get the wrong chip.

That's the sort of improvement you should see in single-threaded performance (based on clock speed alone). If the app is multi-threaded, then you should check that it's utilising both cores in task manager.

 

[21stHermit]

Originally posted by: betasub

Originally posted by: 21stHermit
I just did a bench today on my most important application, its even multi-threaded for dual cores and I got an insignificant 1 min improvement, ~5%.

So far the Athlon 64 X2 4200+ Manchester is a big disappointment.

BTW, in POST it did show up as a Athlon 64 X2 4200+ so I didn't get the wrong chip.

That's the sort of improvement you should see in single-threaded performance (based on clock speed alone). If the app is multi-threaded, then you should check that it's utilising both cores in task manager.

Well, I did open Task Mgr, it showed two cores in the Performance Tab. But given the 1.8 GHz to 2.2 GHz, I would have expected 20% as a minimum improvement.

Am I missing something? I'll happily run another bench.

Thanks
Hermit

 

[GuitarDaddy]

Have you checked CPU usage while running the bench?

Either your app is not using the second core or your I/O limited. If your 15 binaries are writting to disk it's most likely an I/O bottleneck.

 

[htne]

Well, I did open Task Mgr, it showed two cores in the Performance Tab. But given the 1.8 GHz to 2.2 GHz, I would have expected 20% as a minimum improvement.

Am I missing something? I'll happily run another bench.

I expect you have reached the point where the cpu is no longer your limiting factor. Possibly hard disk I/O is now holding you back, just as a guess. You need to watch the Performance Tab in Task Manager while you are running this application. Are both of the cpus at or near 100%? If not, then increasing cpu speed/performance will probably not help that much.

Looking back at your original post, that E1505 is a notebook, right? Probably the reason it did so poorly was because notebooks tend to have slow hard drives. And the Celeron D at 3.46 ghz should have had performance much closer to the Athlon 64. Probably not as good, but much closer that what your numbers show. I suspect that the benchmark numbers you are posting are related more to other factors than to CPU speed/performance.

 

[21stHermit]

Thanks GuitarDaddy & htne,

Benchmark Test:
Compile 60.5MB of text files to 15 finished binary files totaling 10.4MB, this is a 1 x 1 degree extent of 40-foot contour maps.
E1505........ 52 min
CeleD360... 30 min
Athlon64.... 19 min
A 64 X2 .... 18 min

I ran another test of the X2 whereby I moved the files from the 10/100 NAS to the C: drive. I got the same 18 min benchmark. Given the small size of the files, ~5MB for the text files, ~1MB for the binaries, I'm not surprised. It took 10 sec to move all the files from the NAS to C: for the test.

The E1505 laptop was using its internal HD not the NAS, given the above test, I'd expect no performance hit using the NAS. Since each PC has 2GB of RAM, no reason for I/O to be a factor.

Any more ideas?

Hermit