Core 2 vs Piledriver FX?

[Dent1]

Opening this thread to correct some of the misguided information in this community. In light of another thread where a few individuals believed than the Core 2 Duo E8400 could yield better performance in single threaded applications or games than the Piledriver FX 6300.

Some people even claimed that the Phenom II X4 couldn't even compete with the Core 2 Duo.

Upon posing numerous links rebutting this, nobody could challenge it with a logical argument.

This thread gives anyone the chance to prove me wrong.

To kick this thread off I will start with this.


Lets concentrate on the games on Anandtech (http://www.anandtech.com/bench/product/56?vs=80)
E8400 vs Phenom II 940 (both 3GHz)
79.8 vs 76 FPS
45.3 vs 48.2
117 vs 116.2
87FPS vs 82.8FPS

^ above shows the E8400 and Phenom II 940 performing about the same in games. Common sense would say the FX6300 would outperform the E8400 in games given the Piledriver FX's IPC is up to 25% better than its predecessor.

Would this alone be a logical argument? If anyone disagrees I'm happy to rebuttal with more evidence.

 

[Steevo]

All current and prior AMD chips suffer from horrible latency in their caches, thus the need to overclock the on die northbridge to speed up access, and use memory with better timings even at lower frequency than comparable Intel offerings.


I would run a moderate bump in CPU speed, but try to reach as high of a northbridge clock as you could.

 

[mroofie]

Opening this thread to correct some of the misguided information in this community. In light of another thread where a few individuals believed than the Core 2 Duo E8400 could yield better performance in single threaded applications or games than the Piledriver FX 6300.

Upon posing numerous links rebutting this, nobody could challenge it with a logical argument.

This thread gives anyone the chance to prove me wrong.

To kick this thread off I will start with this.


Lets concentrate on the games on Anandtech (http://www.anandtech.com/bench/product/56?vs=80)
E8400 vs Phenom II 940 (both 3GHz)
79.8 vs 76 FPS
45.3 vs 48.2
117 vs 116.2
87FPS vs 82.8FPS

^ above shows the E8400 and Phenom II 940 performing about the same in games. Common sense would say the FX6300 would outperform the E8400 in games given the Piledriver FX's IPC is up to 25% better than its predecessor.

Would this alone be a logical argument? If anyone disagrees I'm happy to rebuttal with more evidence.

core duo challenges a bigger processor
that alone says it all
Its time to stop being faithful to the core my dear friend
It's time to embrace Intel

 

[Dent1]

All current and prior AMD chips suffer from horrible latency in their caches, thus the need to overclock the on die northbridge to speed up access, and use memory with better timings even at lower frequency than comparable Intel offerings.
I would run a moderate bump in CPU speed, but try to reach as high of a northbridge clock as you could.

All current and prior AMD chips suffer from horrible latency in their caches, thus the need to overclock the on die northbridge to speed up access, and use memory with better timings even at lower frequency than comparable Intel offerings.


I would run a moderate bump in CPU speed, but try to reach as high of a northbridge clock as you could.

Please address the evidence in post #1. (Anandtech Review)


I want a constructive debate so we can all learn.

 

[mroofie]

Please address the evidence in post #1. (Anandtech Review)


I want a constructive debate so we can all learn.

is this gonna be another amd fanboy thread? If so I'm out of here =_=

 

[Dent1]

is this gonna be another amd fanboy thread? If so I'm out of here =_=

Yes, because nobody wants to address the evidence in the review. They rather derail it so nobody learns.

I welcome all opinions. I just want a scientific argument proving your case based on a review.

 

[MxPhenom 216]

Yes, because nobody wants to address the evidence in the review. They rather derail it so nobody learns.

I welcome all opinions. I just want a scientific argument proving your case based on a review.

Welcomes all opinions, then argues senselessly everyone's opinions like he is right and they are wrong till no one gives a flying f anymore.

 

[mroofie]

Yes, because nobody wants to address the evidence in the review. They rather derail it so nobody learns.

I welcome all opinions. I just want a scientific argument proving your case based on a review.

for my own sanity and personal being(my soul)
I will not get into a debate or argument
There's so many facts that go against your debate it's pointless

 

[Steevo]

Please address the evidence in post #1. (Anandtech Review)


I want a constructive debate so we can all learn.

Everyone who knows about this understands the issue the last Deneb cores were closer to the C2D in performance per clock than the next generation in its first iterations.


What you are doing comparing older tech is beating a dead horse, and the second issue with that comparison for most was the C2D would reach tremendous clocks on decent voltage and further destroy the AMD performance, I had a 940, and built C2D systems, I pushed hard to reach 3.8Ghz under water, and the C2D's all reached 4Ghz with a minor voltage bump and little else.

You started another thread about why a new X6 AMD was slower than your old system, and it comes down to the per core IPC, AMD is about 25% slower on that generation than Intel. Its why I have kept my 1100T for so long, there is no reason to upgrade yet.

 

[Dent1]

There's so many facts that go against your debate it's pointless

Yet you are unwilling to show these facts?

You can't say the facts go against me but don't produce any facts. It's like saying its a fact the earth is square but don't provide any proof of that claim.

You started another thread about why a new X6 AMD was slower than your old system

Steveo, the last thread wasn't mine. It was a gentleman called Mohammed
http://www.techpowerup.com/forums/threads/low-gaming-performance-fx6300.207591/

 

[MxPhenom 216]

Everyone who knows about this understands the issue the last Deneb cores were closer to the C2D in performance per clock than the next generation in its first iterations.


What you are doing comparing older tech is beating a dead horse, and the second issue with that comparison for most was the C2D would reach tremendous clocks on decent voltage and further destroy the AMD performance, I had a 940, and built C2D systems, I pushed hard to reach 3.8Ghz under water, and the C2D's all reached 4Ghz with a minor voltage bump and little else.

You started another thread about why a new X6 AMD was slower than your old system, and it comes down to the per core IPC, AMD is about 25% slower on that generation than Intel. Its why I have kept my 1100T for so long, there is no reason to upgrade yet.

He didn't start that thread. That was another user. I think OneMoar mistaked that too.

 

[EarthDog]

Opening this thread to correct some of the misguided information in this community. In light of another thread where a few individuals believed than the Core 2 Duo E8400 could yield better performance in single threaded applications...

Would this alone be a logical argument? If anyone disagrees I'm happy to rebuttal with more evidence.

You could start by comparing apples to apples and not relying on extrapolated information 'prove' your point.

I'm not saying I disagree note.

 

[Steevo]

Yet you are unwilling to show these facts?

You can't say the facts go against me but don't produce any facts. It's like saying its a fact the earth is square but don't provide any proof of that claim.




Steveo, the last thread wasn't mine. It was a gentleman called Mohammed
http://www.techpowerup.com/forums/threads/low-gaming-performance-fx6300.207591/

You are right on the thread.

http://www.anandtech.com/bench/product/203?vs=699


Here is the reason. Architecture changes have a significantly larger effect on system performance, AMD essentially made a server chip with more cache built for better performance when the data doesn't branch, games branch based on player decisions and dependencies, the extra latency for flushing the pipeline to get new data in is what hurts them the most.

A benchmark where the data can be spoon fed to the core runs great, games with excellent pre-caching work great as well, but highly volatile games run poorly, the extra on die cache to make up for it means it has to run slower and thus hurts some performance even more. Unless you are capable of understanding the nuances in each architecture and how they effect performance for different code its worthless having a discussion about it, as it akin to trying to describe the flavor of 8 and the smell of purple.

 

[Dent1]

Looking at Bit Tech review from 2009, gaming review.

E8500 (3.16Ghz) vs Phenom II 940 BE (3.1Ghz) - http://www.bit-tech.net/hardware/cpus/2009/01/08/amd-phenom-ii-x4-940-and-920-review/6

Crysis - 37.9FPS vs 36.9FPS
Farcry 2 - 64.2 FPS vs 63.1 FPS
Half Life 2 EP2 - 149.5 vs 155.5

According to those figures, they perform exactly the same. It seems like Bit-tech supports Anandtech's figures too. So the Core 2 architecture and Phenom II architecture seem virtually the same in games.

Logically, it would be impossible for a Piledriver FX to perform worse than the Core 2 Duo and Phenom II given AMD deliberately ensured up to 25% increase after Bulldozer under delivered. Even if they did no refinements it should be about the same.

 

[EarthDog]

Well, to play devil's advocate...

Far cry and crysis are not cpu limited games though...

 

[Dent1]

Well, to play devil's advocate...

Far cry and crysis are not cpu limited games though...

They are popular games, its almost impossible to find a review that excludes those titles.


Gaming Tests: http://www.extremeoverclocking.com/reviews/processors/AMD_Phenom_II_X2_550_10.html

E8400 (3GHz) vs Phenom II X4 940 (3GHz)

Lost Planet: Colonies (low res)
Phenom II X4 940 - 52.00FPS
E8400 - 39.58 FPS

Farcry 2: (low res)
Phenom II X4 940 - 62.17
E8400 - 67.05

Farcry 2: (high es)
Phenom II X4 940 - 51.20
E8400 - 52.97

Crysis Warhead: (low res)
Phenom II 940: 70 FPS
E8400: 85 FPS

Enemy Territory: Quake Wars (low res)
Phenom II X4 940: 156.80 FPS
E8400: 144.00 FPS

Valve particle simulation
Phenom II X4 940: 80 score
E8400: 59.40 score

Valve VRAD Map Build Time
Phenom II X4 940: 138.33 seconds
E8400: 224 seconds

Interesting set of results. I think one distinct message all the review sites (Bit Tech, Anandtech, Extreme Overclocking) share in common is the final iteration of the Phenom II (C3 stepping) could more than keep up with the Core 2 series in gaming and in some situations outperforming it by a fair bit. I want to dispel this stupid myth that says otherwise once and for all.

Using the evidence above it just seems preposterous that people still believe the Core 2 can outperform the FX Piledriver in gaming when we know its well ABOVE the Phenom II.

 

[Pehla]

i would totaly not call my self fan boy of red or green team...,but i cant belive what i read here
do you ppl want to say that core2duo have beter performance then fx6300???i found that realy hard to belive!!as much as i know those procesors(c2d) are made on pentium III architecture
you want to say that pentium III based cpu have beter performance then fx series???

 

[MxPhenom 216]

Ill go back on my original statement in the previous thread. As this bit of comparison between a chip from the stone ages, vs a CPU released in the last couple years is a bit ridiculous. Just understand that when Bulldozer first released (Piledriver brought improvements upon it yes) it could barely beat Phenom II and Core 2 Quad Q9xxx.

In general, current processors AMD vs current Intel on lga1155 and lga1150, Intel will win in about everything. ESPECIALLY Skyrim, which is what I was saying in the original thread, Intel in general, will be better in that particular title than AMD regardless. Same goes with the Total War series. @crazyeyesreaper can enlighten you on that a bit if you want to argue about it. Current AMD chips have lackluster single threaded performance, and I don't think thats disputable. When it comes to multithreaded, AMD chips do pretty well because that is what they were ultimately designed for.

One thing that kills it for me when it comes to CPUs, is that Intel CPUs are able to provide more performance in about everything I do, and for probably everyone else, while also doing it by using half the power that AMD chips do. Now that says a lot, at least too me. It is not about saving on a power bill, less heat, etc. Look at it in an engineering stand point. If you have something that can process work in less time while using less power to do it, that is a win. That is what computing is coming too as well. The desire for more efficient but at the same time offering more processing power.

i would totaly not call my self fan boy of red or green team...,but i cant belive what i read here
do you ppl want to say that core2duo have beter performance then fx6300???i found that realy hard to belive!!as much as i know those procesors(c2d) are made on pentium III architecture
you want to say that pentium III based cpu have beter performance then fx series???

I thought that Core 2 architecture was COMPLETELY new after Intel was getting slapped in the face by AMD with their original Athlon series? Intel then came out with Core 2, and achieved the performance crown. I know that the Pentium D was 2 Pentium 4 taped together on one die.

 

[Steevo]

FX6300 is gimped compared to older X6 AMD processors, as again, it was a server chip knockoff.

 

[Pehla]

if i understand the OP he says that some ppl claim that C2D have beter performance then FX series..,and all he says that it canot be true!!am i missing something else?
nobody is stupid everyone knows that curent ...,danm even the previus core i series is beter performing then AMD flagship... but i think you miss the point the guy wants to say!!

I thought that Core 2 architecture was COMPLETELY new after Intel was getting slapped in the face by AMD

i would never tough of that also...,but i read it few weeks ago on some website..,im not sure if its info is acurate..,and i spent last 20 min searching for that info but cant freakin find it!
but i did find that when C2D came out some PC info gathering software detect it as two pentium 3 xeon cpu´s
i know we are off topic but i just want to share the info...

 

[eidairaman1]

FX6300 is gimped compared to older X6 AMD processors, as again, it was a server chip knockoff.

You're right the FX line are all server Based chips.

 

[Lionheart]

is this gonna be another amd fanboy thread? If so I'm out of here =_=

AMD fanboy thread?

This site is Pro Intel & Nvidia

 

[MT Alex]

Interesting set of results. I think one distinct message all the review sites (Bit Tech, Anandtech, Extreme Overclocking) share in common is the final iteration of the Phenom II (C3 stepping) could more than keep up with the Core 2 series in gaming...

I should hope so, considering that the E8400 was released in January of 2008 and the C3 Phenom IIs came around in November of 2009.

 

[Arjai]

You are right on the thread.

http://www.anandtech.com/bench/product/203?vs=699


Here is the reason. Architecture changes have a significantly larger effect on system performance, AMD essentially made a server chip with more cache built for better performance when the data doesn't branch, games branch based on player decisions and dependencies, the extra latency for flushing the pipeline to get new data in is what hurts them the most.

A benchmark where the data can be spoon fed to the core runs great, games with excellent pre-caching work great as well, but highly volatile games run poorly, the extra on die cache to make up for it means it has to run slower and thus hurts some performance even more. Unless you are capable of understanding the nuances in each architecture and how they effect performance for different code its worthless having a discussion about it, as it akin to trying to describe the flavor of 8 and the smell of purple.

I can explain how I feel when I see purple....I have not smelled it, however, smells can make me see purple. Hmmm, I understand your point but, you may not understand mine...

http://en.wikipedia.org/wiki/Synesthesia
http://www.psychologytoday.com/basics/synesthesia
http://www.apa.org/monitor/mar01/synesthesia.aspx

I mostly feel colors, sometimes hear them, i.e. music. Numbers, not so much, however, formula's sometimes evoke a color. We are not all alike, except that we know what each other means.

 

[HumanSmoke]

I should hope so, considering that the E8400 was released in January of 2008 and the C3 Phenom IIs came around in November of 2009.

Hey, nothing screams relevance like re-re-re-rehashing performance characteristics of 7 year old processors. I heartily look forward to the OPs K7 Argon vs PIII Katmai gaming showdown.

I thought that Core 2 architecture was COMPLETELY new after Intel was getting slapped in the face by AMD with their original Athlon series? Intel then came out with Core 2, and achieved the performance crown. I know that the Pentium D was 2 Pentium 4 taped together on one die.

Kind of a parallel development thing. NetBurst basically played out the string with the Pentium D MCM series, but Core2 was a refinement of P6 and following Core (Centrino/Pentium M) low-voltage architecture developed by Intel's Israeli team which got a boost when Intel shelved the NetBurst based Timna SoC which effectively stopped NetBurst in its tracks R&D wise