P4 OC'd @ 3.5GHz vs A64 3400+

[justly]

"Take this to the PM's....."

After you.

"Go talk to 1,000's of distributed computing ppl and tell them they don't think that is a reason to upgrade...."

If this is their sole reason to upgrade I don't want to talk to them, since they must not have a life

"I did not make that car analogy and cannot be responsible for every persons post in that thread..he is not too far off I found in my testing...i say more like a Z28 with a honda accord (not jacked up!!)...."

I didnt say it was your post, just that you agressivly try to "correct" AMD users but are more lenient about correcting Intel propaganda.

"You go get a p4c w/ HT and dispute any of my benches or claims to date..then you can talk...."

I'm not trying to disprove your benches, am I? ... and if I need to have a P4 w/HT to "dispute" your claims then I would say you need an Athlon 64 before you try to dispute others about its performance.
Whats good for the goose is good for the gander


Rivo, I'm sure both will perform admirably so I would look at the other aspects to make a decission, things like motherboard features, upgradeability, ect... also ask yourself a few questions like what will be more benificail for you HT or 64 bit, what speeds are acceptable if you don't reach your desired overclock on either platform.

 

[RussianSensation]

first overclock 2.8 p4 to 3.5 then ponder whether it's faster than 3400+

You simply cannot make a comparison unless you overclock the 3400+ as well or you are castrating its additional gain as well. If you were gonna go by price performance without overclocking 2.8 vs. 3400+ 2.8 is a better value. However for someone whose income is substantial it doesnt make a difference.

With overclocking, it fair to compare 2.8@3.5 to 3400+ 2.2@2.4 (assuming 10% overclock average for Athlon 64 machine). Even if you run 3400+ at 2.2 it will beat p4 in gaming most of the time. In other applications p4 will dominate it all depends on what programs you use. You can probably see how arguying either side is pointless since many other factors determine what cpu is better.

You can try to estimate the performance difference by reading some reviews like this oneAthlon 64 3400+ vs. the rest and on sites like Anandtech..etc.

It would be just easier to ask someone who has a 3400+ (or 3200 + @2.2ghz) to run some benchmarks and let Duvie with a 3.5ghz p4 run a set of same benchmarks. Also for everyone general usage applications can be defined differently. For some its multitasking and calculating math programs and running SETI, for others its photoshop, and then for others its gaming....these can all be considered general usage if you define "general" as the most common application you run at home...There is little point in arguying this because someone who wants to buy a p4C solution clearly doesnt have money to shell out for an Athlon 64, or possibly has made up his/her made on an Intel solution, or enjoyes overclocking regardless of a small tradeoff in speed in comparison to another solution that's less overclockable. Someone who wants a no hassle free all-around fast solution with a greater budget could be better off by just getting an Athlon 64.

"Go talk to 1,000's of distributed computing ppl and tell them they don't think that is a reason to upgrade...."

And for everyone whose sole purpose to buy a fast computer is to do distributed computing...they should get a life and do something productive like get a university degree for crowing out loud. If you want to help in medicine join the field, and if you want to learn about space become a scientist or a space engineer instead of "trying to search aliens" and "fold proteins". Yes it's important, but only to some extent...If you want to be a helping person then just give $1000 US for that computer you just built to starving childern of the world when 33000 kids die a day of hunger, I bet you'll save at least 1 life. So an argument about how fast your cpu calculates "these" programs is ridiculous as it shows just bragging rights for Oh wow you have the fastest Seti or PIFast times...when those things never affect you in real life. However, real world applications like gaming, decoding, archiving, and office applications are what the majority of the people do use. Let someone benchmark those and then compare.

 

[Duvie]

Originally posted by: justly
"Take this to the PM's....."

After you.

"Go talk to 1,000's of distributed computing ppl and tell them they don't think that is a reason to upgrade...."

If this is their sole reason to upgrade I don't want to talk to them, since they must not have a life

"I did not make that car analogy and cannot be responsible for every persons post in that thread..he is not too far off I found in my testing...i say more like a Z28 with a honda accord (not jacked up!!)...."

I didnt say it was your post, just that you agressivly try to "correct" AMD users but are more lenient about correcting Intel propaganda.

"You go get a p4c w/ HT and dispute any of my benches or claims to date..then you can talk...."

I'm not trying to disprove your benches, am I? ... and if I need to have a P4 w/HT to "dispute" your claims then I would say you need an Athlon 64 before you try to dispute others about its performance.
Whats good for the goose is good for the gander


Rivo, I'm sure both will perform admirably so I would look at the other aspects to make a decission, things like motherboard features, upgradeability, ect... also ask yourself a few questions like what will be more benificail for you HT or 64 bit, what speeds are acceptable if you don't reach your desired overclock on either platform.

that is why I have had Markfw900 and Dapunisher run test for me on their A64's oc'd...I am using there results!!! I can speak with experience of the barton 3200+ in multitasking but I never said I can about the A64...


Edit: I took it to the PMs before your response, you must have chosen not to respond to it.....

 

[Revo]

I can see you dudes take your CPU's very seriously, a bit too seriously if you don't mind me saying! In any event, thanks for your comments. At the moment I'm like a damn yoyo and I'm leaning towards the A64 but I'm taking Jeff's advice and holding back for a while. I think things are going to go 64bit soon especially with the new 64bit WinXP and others will follow suit. As such, I think it would be advantageous to have 64bit capability therefore, an A64 at the right price. I feel we will see a "right" price for the current batch of processors (i.e. P4 Northwood & Socket 754 A64's) in a few months time when the prescott makes it's debut along with maybe new A64's.

From what I've learnt here I would say I will probably go for an A64 3000+ or 3200+ as they can both most probably be OC'd to around 2.2GHz and would then be able to outpace in certain cases and match in many the P4EE which is no mean feat.

EDIT: I initially didn't see your post Russian dude but the comments you made make sense especially regarding the 2.8C being the better option price/performance wise at the moment. However, as the prices drop for the A64, and in particular the 3000+ I feel that it will be the better option price/performance wise. What do you think?

 

[RussianSensation]

Revo, I agree with you on the point that you made regarding the upcoming price drops of the Athlon 64 product line as the 3700+ model number becomes introduced sometime in Q2. However, Intel is going to reduce their prices as well with the 2.8C to be around $178 on February 2nd, along with the rest of their lineup. I dont want to start a flame war of sorts, but I do think if you are simply looking at socket 754 upgrade right now, then for overall performance in your case, which you base on games and general usage + ability to support 64-bit computing, I would generally give an edge to Athlon 64 3000+ over 3.0ghz p4 assuming equal (or close) pricing and similarly to Athlon 64 3200+ over 3.2ghz p4. Preliminary tests do not show Prescott to be performing faster than the Northwood processor at identical clock speeds, thus leading me to believe that prescott (on socket 478) will not be able to overtake the Athlon 64 platform just yet. Since the "revised" p4 is coming in early February, I would wait 10 days to see the benchmarks and evaluate whether it is important for you to upgrade right now, and whether or not the Prescott is a better choice than the Northwood core.

Try not to hypothesize too much about the performance of the upcoming dual channel Athlon 64 on the 939 socket vs. the Prescott processor on Socket T. You could estimate the Athlon to gain a boost of 5% with the dual channel solution giving it a slight edge over similarly clocked FX considering it will have non-registered low latency memory and possibly an added boost from the nforce 3 250 and via pro chipsets. The new 915 pentium chipsets should also provide a slight boost in speed. I think it is impossible to decide right now whether 939 > or < 775 solution, until they are released and fully evaluated based on features, price, overclockability, etc. However, right now the Athlon 64 at around $220-$260 isn't such a bad deal vs. the 478 platform considering it reaches its end of the line with 3.4ghz, but the Athlon will be upgradable to at least 3700+ (which isnt a lot of room either).

Personally, I think you should wait just a bit longer and wait for the new sockets for each platform and assess their price performance ratio then. In either case, both of the new platforms will provide you with an extensive upgrade path, and you could argue that if the perfomance today isnt satisfactory with the 3xxx CPUs you always can simply buy a faster processor without having to consider a new motherboard.

 

[digits]

I cannot add much here but some FACTS. I and another person On a different forum did some 3dmark tests with 2 setups mine a P4 2.8 @3.5 and his an A643000+ @2300 with the same type of vid card ( using same drivers, and stock speeds ) Are results came out to be about the same or within 50 points of each other.

So there Ya go, 2 system you asked about In a gaming enviroment ( or at least a gaming enviroment ) and both with what I call average overclocks.

 

[Revo]

I was just wondering, and I'm slightly digressing, when OC'ing an A64, or any other CPU for that matter, is the memory divider pre-set or can you set it to requisite values (i.e. any ratio)? The reason I'm asking is that I have some Crucial PC3200 which I intend to use in my new system. If I opt for the A64 3000+ or 3200+ then the max OC I'm looking at is roughly around 2.2 - 2.4GHz which will be set by a 'FSB' of 220 or 240MHz. As such I will have to use the memory divider to successfully OC the A64 and if the memory dividers are pre-set then I may have to run the memory at a frequency lower than its maximum. For example, if I use a 'FSB' of 220MHz and a 5:4 divider then my memory will be running at 176MHz. This in turn would reduce the performance somewhat in which case I would probably be better off with a P4. So are the dividers pre-set?

 

[Revo]

Much appreciated Digits! Can you provide the link for that comparison?

 

[digits]

I personaly have never seen dividers be flexible. only 5/4 , or 3/2 , or even 4/5 for older systems. You would have same type of problem on a p4 system if it does not get to 3.5 also. you would either just take the hit or buy ram to run 1:1 with what ever setup you decide on.

 

[ntrights]

Not sure about the 2.8 but ive seen 3's do 3.5+ easy with air cooling..
pc3200 should be ok, use a divider and try diffrent timings and see what works best for your system.

 

[Revo]

Has anyone known a P4 2.8C not to get up to 3.5GHz+?

 

[digits]

getting to 3.5 on a p4 isn't cake like manny seem to think so yes there are a few that dont make it.
also check your messages revo

 

[Revo]

Messages? Which, the ones I post here?

EDIT: It's OK I figured it out!

What even the retail version have difficulty in getting up to 3.5GHz?

 

[PCTweaker5]

Wow Revo! Your hair must be grey now after all this stress and arguements. lol jk have a nice day

 

[digits]

What even the retail version have difficulty in getting up to 3.5GHz?

There are no garauntee's in overclocking. My 2.8 @ 3.5 requires 1.65v and others do it with a lot less. It's just hit and miss. Also the rest of the system gotta handle what you want it to do or the cpu speed will suffer.

Your best bet is to always expect 2.8 from a 2.8 and ya cant go wrong. The same goes for a 3000+

 

[BugsBunny1078]

The benchmarks I have seen show that the p4 3.2 ghz beats the athlon 64 3200 in every benchmark.
Really I resent the 3200 rating as it is a scam trying to fool people into thinking they have a 3.2 ghz processor when they do not.I do know alot of people who had athlons and thought that 2600 meant 2.6 ghz. I wouldnt buy another pr rated processor on purpose after getting scammed by cyrix and radio shack on the pr166.

 

[PetNorth]

The benchmarks I have seen show that the p4 3.2 ghz beats the athlon 64 3200 in every benchmark.

Sure?

look at this:

http://www.xbitlabs.com/articles/cpu/display/athlon64-3400.html

Here you have A64 3200+ and P4C 3.2

 

[justly]

BugsBunny1078, sounds like you need to read a little more, you can start off with the
link PetNorth provided

Also, many people share your concerns about the pr numbers due to how different benchmarks favor different processors, and the great performance leap the P4s see in comparison when its FSB is increased.

The thing is, using GHz by itself is just as flawed.

Don't believe me ...do you Are you aware that newegg carries six different Intel 2.4GHz processors varying in FSB speeds, HT capabilities and cache sizes. want to make a bet that they don't have the same performance

Duvie, I hope I didn't blow this one out of proportion