Equivalent athlon speed compared to duron 1.6ghz?

[blazer78]

hi i was just wondering how fast an athlon would have to be to compete against a duron 1.6 ghz on the new applebred core. thnx

(my estimate is athlon 1.1ghz =) i'm sure i'm wrong though..)

 

[Duvie]

That is tough to gauge since not all programs react the same to the added l2 and larger l1 cache of the athlons.....

other then cheap home theater boxes and ultra cheap interet surfer computers I can't see why anyone buys these neutered athlon xps.....


I am not a 100% familiar with this new version so maybe list what the l1 and l2 cache specs are as well as the fsb, please.....

 

[blazer78]

umm.. i think 128kb L1 and 64kb L2 , although i read somewhere that u could unlock the duron so that it has the same amound of cache as a thoroughbred B would..

correct me if i'm wrong..

 

[SickBeast]

Originally posted by: blazer78
umm.. i think 128kb L1 and 64kb L2 , although i read somewhere that u could unlock the duron so that it has the same amound of cache as a thoroughbred B would..

correct me if i'm wrong..

There may be a way to unlock it, but there is always the chance that the chip was castrated for a reason...I.E. some of the cache is defective and doesn't work. If what you're saying is correct and the Durons only have 64kb of L2 cache, I would highly reccomend you AVOID buying them. The Barton XPs have 512kb of L2...that's 8 times as much!

To answer your question, I would say that a vanilla Athlon 1.2ghz would compare to the Duron at 1.6.

 

[pyrojunkie]

Durons have 128kb of L2 cache and 64K of L1 cache. All XP processors have 64kb of L1 cache.

 

[MDE]

Originally posted by: pyrojunkie
Durons have 128kb of L2 cache and 64K of L1 cache. All XP processors have 64kb of L1 cache.

No, the Durons have 128k L1 and 64k L2.

 

[SickBeast]

Originally posted by: pyrojunkie
Durons have 128kb of L2 cache and 64K of L1 cache. All XP processors have 64kb of L1 cache.

All XP processors have 128kb of L1 cache.

 

[blazer78]

lol, so a athlon 1.2 .... hmm... thats not that bad i spose...

 

[justly]

I would have to be a little more optimistic.

Some apps are going to see a bigger hit than others, but the change from 256KB to 512KB L2 cache had a max gain of 8.2% in non synthetic benchmarks.

AMD's Athlon XP 3000+: Barton cuts it close

The average was around 5 to 6% in most games and considerably less in other apps.

So if you take 10% away from an equally clocked Athlon XP (1900+) it comes out to be closer to a Athlon Xp 1600+.

All in all I would expect performance in the range from a Athlon XP 1500+ to 1900+ depending on how cache intensive the app is.

 

[SickBeast]

Originally posted by: justly
I would have to be a little more optimistic.

Some apps are going to see a bigger hit than others, but the change from 256KB to 512KB L2 cache had a max gain of 8.2% in non synthetic benchmarks.

AMD's Athlon XP 3000+: Barton cuts it close

The average was around 5 to 6% in most games and considerably less in other apps.

So if you take 10% away from an equally clocked Athlon XP (1900+) it comes out to be closer to a Athlon Xp 1600+.

All in all I would expect performance in the range from a Athlon XP 1500+ to 1900+ depending on how cache intensive the app is.

Dude, we're talking about 64kb L2 DURONS here.

 

[InlineFive]

Wait, I thought the Durons had 128KB L1 and 64KB L2? Or is that the other way around now?

 

[SickBeast]

Originally posted by: PorBleemo
Wait, I thought the Durons had 128KB L1 and 64KB L2? Or is that the other way around now?

Thou art correctamundo.

 

[justly]

Thats right we are talking about Durons.

If you think I am wrong provide some feedback why your estimate of a 1.2GHz Athlon is more acurate than my estimate of a 1600+ Athlon XP, and please feel free to provide any links to prove your point.

 

[SickBeast]

Originally posted by: justly
Thats right we are talking about Durons.

If you think I am wrong provide some feedback why your estimate of a 1.2GHz Athlon is more acurate than my estimate of a 1600+ Athlon XP, and please feel free to provide any links to prove your point.

UGH...I posted a link to it on here the other day. It's in the huge CPU roundup at Tom's Hardware.

In your quote you took 10% off the performance of an Athlon XP, but you were taking your figures from an article comparing a Barton with a Tbred-B! Saying a Duron 1600 is equivalent to an Athlon XP 1600+ is like telling people that a 2.0ghz Celeron is just as fast as a 2.0ghz P4 w/800mhz FSB. Sorry...it does not compute.

 

[cow123]

i dont think L2 cache is all that important in athlons anyway. i've disabled level2 on my barton chip and actually barely noticed a difference

 

[SickBeast]

Originally posted by: cow123
i dont think L2 cache is all that important in athlons anyway. i've disabled level2 on my barton chip and actually barely noticed a difference

Uhhh...yeah...I really don't know what to say to that other than...yah...ok...sure...whatever you say.

Try a benchmark with your cache disabled.

 

[blazer78]

Originally posted by: SickBeast

Originally posted by: justly
Thats right we are talking about Durons.

If you think I am wrong provide some feedback why your estimate of a 1.2GHz Athlon is more acurate than my estimate of a 1600+ Athlon XP, and please feel free to provide any links to prove your point.

UGH...I posted a link to it on here the other day. It's in the huge CPU roundup at Tom's Hardware.

In your quote you took 10% off the performance of an Athlon XP, but you were taking your figures from an article comparing a Barton with a Tbred-B! Saying a Duron 1600 is equivalent to an Athlon XP 1600+ is like telling people that a 2.0ghz Celeron is just as fast as a 2.0ghz P4 w/800mhz FSB. Sorry...it does not compute.

i thought tomshardware was biased since intel is a sponsor for tomsharware.... :s

 

[dguy6789]

a duron 1.6 will be around as fast as an athlon 1 - 1.2 ghz(that is athlon origonal, the athlon xp would be more so) i know this because i have had both, and the athlon 1 ghz oced to 1.2 was faster in everything, and the duron benched between 1 ghz and 1.2 compared to the athlon.

 

[justly]

Blazer78, you could say that ALL sites have some amount of errors or odd judgement calls on what benchmarks to use, and that will make any of them look biased to a degree. Remember that different benchmarks are going to favor different aspects of a CPU design so just be aware of your needs and base performance on benchmarks that fit your needs, and a lot of that bias will be easy to avoid.

SickBeast, as far as I am aware TomsHardware hasn't done a CPU roundup since the Applebred core was introduced, are you sure the Duron in that roundup wasn't the Morgan core.
Well, I just did a search for "Duron" at Toms and the most recent article that came up was Benchmark Marathon: 65 CPUs from 100 MHz to 3066 MHz in this article the fastest Duron was the Morgan core at 1.3GHz and it normally ran near the 1.2 Athlon "B" with the same memory. The Duron 1.3GHz actually benchmarked faster than a Athlon "C" 1.33GHz with faster memory in 1 benchmark and slower than the Athlon "B" 1.0GHz with the same memory in 2 benchmarks, but those three exceptions where all synthetic benchmarks.

Based on this information (if memory and FSB is equal), we are expected to believe the added 300MHz speed increase and faster FSB that the 1.6GHz Duron has would have no effect on performance...I would say that doesn't compute.

If you want to say that I am to optimistic then I think you should also conceed that you may be to pesimistic and maybe the truth lies in the middle somewhere.

I will grant you one thing, newer games being run at higher resolutions will demand more bandwidth and this is where the Duron will probably perform the worst when compared to the Athlon.

 

[Assimilator1]

Originally posted by: SickBeast

Originally posted by: blazer78
umm.. i think 128kb L1 and 64kb L2 , although i read somewhere that u could unlock the duron so that it has the same amound of cache as a thoroughbred B would..

correct me if i'm wrong..

There may be a way to unlock it, but there is always the chance that the chip was castrated for a reason...I.E. some of the cache is defective and doesn't work. If what you're saying is correct and the Durons only have 64kb of L2 cache, I would highly reccomend you AVOID buying them. The Barton XPs have 512kb of L2...that's 8 times as much!
.

Jez some of you guys have a narrow field of vision! ,you are comparing a CPU that costs about £34 to a cpu that costs £65 (min) ,when your building a PC purely for Internet/office type use that extra cache & performance is simply a waste of money.

 

[blazer78]

lol, how bout a duron enthusiast?

 

[aka1nas]

Also, the original applebred threads that I saw at (I think) overclocker's reported that
1) Most of the Applebreds would unlock to full t-breds easily and stably
2) The performance penalty for having less cache was on the order of ~<15% for most apps
3) The chip seemed to overclock much better at a given voltage, prob due to fewer transistors being operational due to the decreased cache.

Keep in mind that the barton is now the high end core and thorton(cut down barton) is replacing t-bred as the lower end chip. They are probably trying to sell through that inventory of t-breds as quick as possible so they only have 1 core design to worry about, hence applebred.

 

[sandorski]

The Duron certainly doesn't suck compared to Athlon XPs the way a Celeron sucks compared to P4s. They are "crippled", but respectable, especially for the price.

 

[blazer78]

y do i keep hearing that the P4 sucks? i heard the older models of the P4 couldn't even compete with durons, i dun think thats true or is it?

 

[Assimilator1]

Actually he was referring to P4 based Celerons 'sucking' compared to true P4's ,there is a huge difference in performance,the P4 Celerons suck bigtime ,e.g. a PIII 1.4GHz is much faster than a Celeron 2GHz!.

Oh yeah ,& the Duron gave the early P4 (willamette) a good run for its money!
A PIII 1GHz was about as fast as a P4 1.4GHz