Sign in with
Sign up | Sign in
Your question
Closed

Core2Quad Q9705 3.16 GHz vs. Core i5 3450 3.1 Ghz... which is faster?

Last response: in CPUs
Share
January 19, 2013 3:39:46 PM


I had a core 2 duo e7500 which i upgraded to i5 3450 (after consulting users of tomshardware.com - and thank you guys, the processor is zillion times faster than e7500) but what surprised me while i searched all over web for suitable cpu options, that processors with same speed are so varied in prices and even generation after generations of innovation doesn't guarantees in giant leap in speed as far as Ghz for cpu goes, so my question is...

as i understand Ghz is Ghz, and no matter what generation/model/ architecture etc. cpu belongs to it will be same as far as calculations per clock cycle goes (i hope i am saying it right!). Now then, how is a Core i5 3450 3.1 Ghz faster then a Core2Quad Q9705 3.16 GHz. in this case even cores are equal in number. Even certain sandy bridge's certain 2.x chips are faster than my cpu. Without overclocking (like 2600k) is the difference that significant.


If i throw the additional sizzling gimmicks like HD 2500 and HD 4000, the company is recycling same quality at higher rates. What is the point of producing cpu after cpu if the speed doesn't go from 3 to 6 in next generation, the 32nm to 22nm and now to 14nm... why so much waste of dye surface if there is no significant addition in performance.

Don't tell me if i am a troll, i already know. I am sparking a conversational topic, who knows we come up with some amazing realizations here.
a c 83 à CPUs
January 19, 2013 4:37:53 PM

Clock speed comparisons only work within the same CPU architecture. Every generation of processors from Intel is a new and improved architecture, capable of processing more instructions per cycle meaning they do more at the same clock speed.

Intel learned with the Pentium 4 there are diminishing returns for continued increase in clock speed. The more you increase clock speed the more voltage the chip requires, and the more power it draws and heat it produces.
Score
0
Related resources
January 19, 2013 4:49:43 PM

13thmonkey said:
performance isn't measured in GHz, do some research. It's all about instructions per cycle.
http://www.anandtech.com/bench/Product/49?vs=363
I5 2400 is slower than the 3450 and 9650 is a little slower than the 9705.



2500K is faster than i5 3450, what is the point of taking one step ahead in technology, when there is no improvemnet in speed. Sandybridge model beat Ivy Bridge??? That's not fair.
Score
0
January 19, 2013 4:52:39 PM

loneninja said:
Clock speed comparisons only work within the same CPU architecture. Every generation of processors from Intel is a new and improved architecture, capable of processing more instructions per cycle meaning they do more at the same clock speed.

Intel learned with the Pentium 4 there are diminishing returns for continued increase in clock speed. The more you increase clock speed the more voltage the chip requires, and the more power it draws and heat it produces.



what i meant was, shouldn't they make some criteria that clearly explains a consumer how much more fast one chip is from other, 3.x in two different quads look the same to any normal person.

Score
0
a b à CPUs
January 19, 2013 4:58:12 PM

Gentlemen. Simple enough. Core i5 3450 has a better architecture which allows the instructions per clock to be much better. GHz doesn't matter these days unless we're deciding who's 3570K is faster (overclocking). The reason why the 2500K is faster is because it has a higher clockspeed, which cancels out the 15% performance increase with Ivy Bridge. Now, if we compared the 2500K to the 3570K, you'll see that the latter is faster. With GHz, it doesn't show which chip is faster, it just shows how many instructions are done in a cycle.
Score
0
January 19, 2013 5:04:28 PM

Is there any article that explains exactly what i am asking here. I have a i5 3450 and i have seen how a Quad core works, so i know first hand... the all performance increase business. I want to know the science of it, that's what matters. As a consumer... i am saying this for folks who buy core 2 duo @ 3.x Ghz believing its the same as i3 two core at the same speed.
Score
0
January 19, 2013 5:08:42 PM

for example once a dear friend of mine came back fuming at home, thinking the seller was duping him into buying expensive new cpu when it was all...

as he said 'core 2 duo is i3, core 2 quad is i5 and i7 (the six core) was the combo of the two' that is what it looks to a ignorant consumer.
Score
0
a c 79 à CPUs
January 19, 2013 5:48:40 PM

austenwhd said:
2500K is faster than i5 3450, what is the point of taking one step ahead in technology, when there is no improvemnet in speed. Sandybridge model beat Ivy Bridge??? That's not fair.


boo hoo its not fair.

ffs

the 2500K is the top i5 the 3450 is not the top i5, the 3570K beats the 2500K by 5-10% for the same clock speed, ergo it has gotten faster.
Score
0
a c 79 à CPUs
January 19, 2013 5:51:44 PM

austenwhd said:
for example once a dear friend of mine came back fuming at home, thinking the seller was duping him into buying expensive new cpu when it was all...

as he said 'core 2 duo is i3, core 2 quad is i5 and i7 (the six core) was the combo of the two' that is what it looks to a ignorant consumer.


i7 is not six core.

simply put every generation will get either quicker for similar power or just slightly quicker but use less power. My Q9550 was about 60% of my 3570K in terms of the performance, and used about 50% more power.
Score
0
January 20, 2013 11:02:09 AM

13thmonkey said:
i7 is not six core.

simply put every generation will get either quicker for similar power or just slightly quicker but use less power. My Q9550 was about 60% of my 3570K in terms of the performance, and used about 50% more power.


Core i7-3960X, Core i7-3930K Processor, Core i7 990X Processor are all six core...

does that show i have done research? i know what i am talking about, 'instructions per cycle' something like a IPC value should be the advertised criteria for selling CPU and not the Ghz, its misleading. All i am asking is somebody to provide me link to an article that explains it in 'English'.


Score
0
January 20, 2013 11:20:28 AM

I get where you are coming from.

More than a few times we have priced up a build for a customer only for them to come back claiming to have found something elsewhere which is faster and cheaper. It turns out they have seen something like an fx-4100 build compared to our machine using an i5-3450.

Of course on here we all know the i5 wipes the floor in performance but in the eyes of a novice:
- both chips are quad-core
- AMD runs at 3.6Ghz
- Intel only runs at 3.1Ghz
- AMD is clearly faster!.... sigh

I have fun trying to explain this to people on a weekly basis :)  Along with graphics cards... average Joe sees a 2GB card and in his mind it's automatically better than any 1GB card in existence because it has double the memory.
Score
0
a c 79 à CPUs
January 20, 2013 11:32:29 AM

austenwhd said:
Core i7-3960X, Core i7-3930K Processor, Core i7 990X Processor are all six core...

does that show i have done research? i know what i am talking about, 'instructions per cycle' something like a IPC value should be the advertised criteria for selling CPU and not the Ghz, its misleading. All i am asking is somebody to provide me link to an article that explains it in 'English'.


The majority of i7s are 4 core + HT, i7 3770K etc.
Score
0
January 20, 2013 11:49:16 AM

13thmonkey said:
The majority of i7s are 4 core + HT, i7 3770K etc.


I don't think anyone is trying to say otherwise.

Some i7's are 4 core + HT (eg. 3770k)
Some i7's are 6 core + HT (eg. 3930k)
The one he was referring to just happened to be 6 core.
Score
0
January 20, 2013 1:32:00 PM

casper1973 said:
I get where you are coming from.

More than a few times we have priced up a build for a customer only for them to come back claiming to have found something elsewhere which is faster and cheaper. It turns out they have seen something like an fx-4100 build compared to our machine using an i5-3450.

Of course on here we all know the i5 wipes the floor in performance but in the eyes of a novice:
- both chips are quad-core
- AMD runs at 3.6Ghz
- Intel only runs at 3.1Ghz
- AMD is clearly faster!.... sigh

I have fun trying to explain this to people on a weekly basis :)  Along with graphics cards... average Joe sees a 2GB card and in his mind it's automatically better than any 1GB card in existence because it has double the memory.


casper1973, i am not one of those people you are talking about, trust me, i am not, if you read the entire thread, you will realize, i am only pointing at the fact that cpu manufacturing companies understand the common people mentality so even if they couldn't get the number to look like 10 Ghz for a processor, which trust me will look good next to a dual core to a 'novice'. I am saying isn't it better to use IPC (Instructions per cycle) instead of Ghz. I am sure a faster processor will have more IPC, no matter what year/ brand/ architecture or Nano Meter it is based on. So instead, they make a 'next' generation model, that runs at 3.x Ghz, and people are like WTH mine runs at 4.x Ghz. I think if IPC becomes the standard, nobody will ever need to go to a sites and check benchmark to know what CPu is fastest. True story!


You say a cheetah runs at 140 Kmps, misleadingly impressive, do you really think he runs all the way to a hundred and forty kilometer to prove that.

Score
0
January 20, 2013 1:35:33 PM

'Core2Quad Q9705 3.16 GHz vs. Core i5 3450 3.1 Ghz... which is faster?' i Know the title is misleading but i will appreciate if you read this thread before commenting. Informative posts expected.
Score
0
a c 471 à CPUs
January 20, 2013 1:38:26 PM

Assuming the same clockspeed an Ivy Bridge Core i5 CPU is about 30% more powerful than a Core 2 Quad CPU.

If you are doing financial or scientific modelling (statistics) then the increase in performance should be even higher because of the AVX (Advanced Vector Extensions) instruction set which more or less can theoretically double floating point calculation performance.
Score
0
January 20, 2013 1:45:15 PM

jaguarskx said:
Assuming the same clockspeed an Ivy Bridge Core i5 CPU is about 30% more powerful than a Core 2 Quad CPU.

If you are doing financial or scientific modelling (statistics) then the increase in performance should be even higher because of the AVX (Advanced Vector Extensions) instruction set which more or less can theoretically double floating point calculation performance.



Read the question's description, please, this is not what i am here for. I am asking for information on the topic, not asking which one is faster. I made a mistake not pointing it out earlier, instead, replied to replies unrelated to my query.
Score
0
January 20, 2013 7:47:09 PM

I wasn't trying to say you were one of those people, just giving an example of how this problem effects me on almost a daily basis. Sorry if it came across otherwise.

I suppose you could have a figure where the IPC is multiplied by the clock speed, and then by the number of cores. But even that wouldn't be accurate. I done the calculations and an FX-6100 thrashes an i5-3450 according to the numbers... which in reality is quite the opposite.

So I'm no further forward with a solution! But I do agree something should be done.
Score
0
a c 79 à CPUs
January 20, 2013 8:10:43 PM

we've given you information, yes perhaps rating a processor at ipc x Ghz might be nice, but what about hyper threading, what about AMD's weird shared functionality between cores, sometimes is great sometimes not. IPC is not a clean measure.
Score
0
January 20, 2013 11:39:30 PM

It's actually the power consumption / heating that they are more concerened about. I think it is not by pumping the clock speed to 5Ghz or higher that makes the difference, rather the architecture itself. Just like the transition from a single core CPU to dual then to Quad Core. It is more important to process threads at the same time rather than processing them faster. ..Having a Faster FSB( although they are not using FSB anymore) then clock speed.

Overall, I think it has evolved. That it is not just Ghz that is important.
Score
0
January 21, 2013 5:20:53 PM

Okay something like 'n bits of bite processed in a n of time' there should be a standard of universal performance benchmark, its like saying there are two types of forces, one is gravity and the other is magnetism... duh!
Score
0
January 21, 2013 5:29:28 PM

payturr said:
Gentlemen. Simple enough. Core i5 3450 has a better architecture which allows the instructions per clock to be much better. GHz doesn't matter these days unless we're deciding who's 3570K is faster (overclocking). The reason why the 2500K is faster is because it has a higher clockspeed, which cancels out the 15% performance increase with Ivy Bridge. Now, if we compared the 2500K to the 3570K, you'll see that the latter is faster. With GHz, it doesn't show which chip is faster, it just shows how many instructions are done in a cycle.


Don't forget the 50% increase in heat production!
Score
0
a c 79 à CPUs
January 21, 2013 5:30:32 PM

but it depends on the type of bytes its processing... floating point vs non-floating point, vs encode decode vs compression, all give different answers. thats why there are suites of benchmarks.
Score
0
a c 79 à CPUs
January 21, 2013 5:31:15 PM

taxemicshadow said:
Don't forget the 50% increase in heat production!


???
Score
0
January 21, 2013 5:40:16 PM

13thmonkey said:
???

It's a joke. For the most part though the Ivy Bridge CPU's run hotter than their Sandy Bridge counterpart.
Score
0
January 21, 2013 10:45:42 PM

Wait, lets say binary of bits, 0,1... i hope 0's and 1's calculated per second is going to be the same with any cpu. processing the same data on two different cpu can tell how many binary data did one did at 'n' given time.

I can imagine there must be some inside industrial standard for benchmarking, a secret.
Score
0
January 21, 2013 10:48:26 PM

13thmonkey said:
???


my i5 3450, runs at 58c, at 100% load, core 0,1,3 runs at 50 and core 2 runs at 61, isn't it wierd, it is OCing at 3.5 Ghz.
Score
0
January 24, 2013 8:56:57 PM

So this thread is pretty much dead. On an end note, i must say this that Intel is overpricing its component, in absence of competition from AMD, basically, they run the market. They pay companies to write code that are optimized for Intel CPU. Also the fact that when you refer to sites like Passmark or Tomshardware for benchmark they start talking about 3DMark and Gaming performance of a CPU. Seriously, this fact is solidly placed that even most extremely demanding modern games don't utilize more than 4 cores (even 2 cores are enough - more than!). What doesn't make sense is the absence of video encoding benchmark... simply saying, if you have a four core AMD (32 nm) or Intel Sandy Bridge, you have more than enough firepower to play any game, processor can only take you as far as getting through the minimum requirement, now all you need is a decent DX11 compatible graphics card, more than enough, or go for a overkill and get a DDR3 ram 1600 instead of 1333.

AMD is cheap, high on TDP, good fast. Intel is expensive, less on TDP, good fast. Remember it's what you want and how much you can spend that matters. At given price range a 100 $ Intel CPU is just as good (sometimes not so) as 100 $ AMD! True story.

Goodbye.

(if you think my English is bad, that's cause i am too lazy to look up a dicktionary)
Score
0
January 25, 2013 1:15:00 AM

austenwhd said:
So this thread is pretty much dead. On an end note, i must say this that Intel is overpricing its component, in absence of competition from AMD, basically, they run the market. They pay companies to write code that are optimized for Intel CPU. Also the fact that when you refer to sites like Passmark or Tomshardware for benchmark they start talking about 3DMark and Gaming performance of a CPU. Seriously, this fact is solidly placed that even most extremely demanding modern games don't utilize more than 4 cores (even 2 cores are enough - more than!). What doesn't make sense is the absence of video encoding benchmark... simply saying, if you have a four core AMD (32 nm) or Intel Sandy Bridge, you have more than enough firepower to play any game, processor can only take you as far as getting through the minimum requirement, now all you need is a decent DX11 compatible graphics card, more than enough, or go for a overkill and get a DDR3 ram 1600 instead of 1333.

AMD is cheap, high on TDP, good fast. Intel is expensive, less on TDP, good fast. Remember it's what you want and how much you can spend that matters. At given price range a 100 $ Intel CPU is just as good (sometimes not so) as 100 $ AMD! True story.

Goodbye.

(if you think my English is bad, that's cause i am too lazy to look up a dictionary)

*slow clap*
Score
0
a c 79 à CPUs
January 25, 2013 6:40:16 AM

austenwhd said:
So this thread is pretty much dead. On an end note, i must say this that Intel is overpricing its component, in absence of competition from AMD, basically, they run the market.


odd how prices dropped per performance point year on year and dropped significantly when core 2 duo came out.
Score
0

Best solution

a c 79 à CPUs
January 25, 2013 7:31:35 AM

austenwhd said:
I had a core 2 duo e7500 which i upgraded to i5 3450 (after consulting users of tomshardware.com - and thank you guys, the processor is zillion times faster than e7500) but what surprised me while i searched all over web for suitable cpu options, that processors with same speed are so varied in prices and even generation after generations of innovation doesn't guarantees in giant leap in speed as far as Ghz for cpu goes, so my question is...

as i understand Ghz is Ghz, and no matter what generation/model/ architecture etc. cpu belongs to it will be same as far as calculations per clock cycle goes (i hope i am saying it right!). Now then, how is a Core i5 3450 3.1 Ghz faster then a Core2Quad Q9705 3.16 GHz. in this case even cores are equal in number. Even certain sandy bridge's certain 2.x chips are faster than my cpu. Without overclocking (like 2600k) is the difference that significant.


If i throw the additional sizzling gimmicks like HD 2500 and HD 4000, the company is recycling same quality at higher rates. What is the point of producing cpu after cpu if the speed doesn't go from 3 to 6 in next generation, the 32nm to 22nm and now to 14nm... why so much waste of dye surface if there is no significant addition in performance.

Don't tell me if i am a troll, i already know. I am sparking a conversational topic, who knows we come up with some amazing realizations here.

GHz is the speed of tabulating an instruction. As architecture improves, instruction handling improves. To understand instructions, consider this hypothetical example: Compute the result of 5x5. to execute this, you need to perform 5+5+5+5+5 right?
now suppose cpu architecture allows me three variables only as A,B and C. so I can store a value in A, then store another value in B then invoke the operator "+" and store the result in C.

to perform a simple task like 5+5+5+5+5, We have to use the above idea as below
A=5, B=5, A(5)+B(5)= C(10)
"clear" A
make C(10)=A, so A=10 now
Clear C, So C is empty again

now we have
A(10)+B(5)= C(15)
.
.
.
.
so on till
A(20)+ B(5)= C(25)
STOP

So we perform three steps of clearing A, performing C=A and then clearing C before each addition cycle.
we do this thrice to get to 25 so we actually end up performing 9 operations extra other than plain addition.

Now suppose a new cpu architecture allows you 6 variables
you just perform
A=B=C=D=E=5
A+B+C+D+E=F
STOP

while we used 13-14 steps using the first cpu's logic, our second cpu allowed us to write a shorter program which will be computed with less instructions as well since you are not clearing and reloading Values in register here.

so even though both CPU's have the same clock speed, and they both execute instructions at the same speed too, second cpu will be faster simply because it has less instructions to execute.

I hope this clears your doubt

good luck
-satyam

p.s. I know any programmer who sees this will probably want to kill me :) 
Share
January 25, 2013 10:59:30 AM

satyamdubey said:
GHz is the speed of tabulating an instruction. As architecture improves, instruction handling improves. To understand instructions, consider this hypothetical example: Compute the result of 5x5. to execute this, you need to perform 5+5+5+5+5 right?
now suppose cpu architecture allows me three variables only as A,B and C. so I can store a value in A, then store another value in B then invoke the operator "+" and store the result in C.

to perform a simple task like 5+5+5+5+5, We have to use the above idea as below
A=5, B=5, A(5)+B(5)= C(10)
"clear" A
make C(10)=A, so A=10 now
Clear C, So C is empty again

now we have
A(10)+B(5)= C(15)
.
.
.
.
so on till
A(20)+ B(5)= C(25)
STOP

So we perform three steps of clearing A, performing C=A and then clearing C before each addition cycle.
we do this thrice to get to 25 so we actually end up performing 9 operations extra other than plain addition.

Now suppose a new cpu architecture allows you 6 variables
you just perform
A=B=C=D=E=5
A+B+C+D+E=F
STOP

while we used 13-14 steps using the first cpu's logic, our second cpu allowed us to write a shorter program which will be computed with less instructions as well since you are not clearing and reloading Values in register here.

so even though both CPU's have the same clock speed, and they both execute instructions at the same speed too, second cpu will be faster simply because it has less instructions to execute.

I hope this clears your doubt

good luck
-satyam

p.s. I know any programmer who sees this will probably want to kill me :) 


I had a friend named Satyam Dubey, (broke his wrist fist fighting, once!) anyway, thanks for making that elaborate explanation. Now that is as simple as it gets answer i was looking for. Thank you. And that brings me to other part of mystery - 'Why Intel specs come overpriced? Is it worth to buy AMD instead?'. In my opinion, yes, a 8 core 8350 at 150 $ (in my country) is worth every single dollar its asks, Intel doesn't. Over i5, for gaming and encoding, every thing else is overpriced and overkill.

(if you think my English is bad, that's cause i am too lazy to look up a dicktionary)
Score
0
a c 79 à CPUs
January 25, 2013 12:29:38 PM

well i have indeed broken quiet a few bones but never my wrist :)  Intel can over price their cpu's because they have been performing better than AMD counter parts and that is in synthetic benchmarks. Not just that, Intel parts at the same price points as AMD's offer good performnce at lower TDP as well so they are power efficient.

AMD compensates that by increasing the default clock of their cpu's and pricing them lower.

all this is a result of the way the two companies have revised their architectures over time. AMD introduced on-die memory controller and 64 bit x86 architecture and were leaders and Intel trailed. Now Intel is leading at the moment and AMD is a little behind.

I think there is a big marketing mistake AMD made with Bulldozer architecture. Intel has had hyperthreading for quiet some time. and you know that a dual core hyperthreaded intel processor looks like 4 cores(two physical/two virtual) to the OS or application. However each virtual core is like .2-.25 of a real core.

AMD saw hyperthreading and actually made a better and faster solution by designing Modules. Each module in bulldozer has two physical cores which shared decoder and cache. this made each core .75 times of a tru core.

So while one intel core could do 1.2 times the job, each AMD module can do 1.75 times the job. Till here ther are clear winners but they made the mistake (atleast I believe so) of calling ther processors Octa-core. Had they just called them Quad-Module or something, their processors would not have had to compete for top place with 6 core i7's. An 8350 is almost as good as a quad i7 and cost's half and beats every i5.

For 150, it is a great chip. In steam roller, the module efficiency would be improved further.
Score
0
a b à CPUs
January 25, 2013 1:33:05 PM

^^ You need to add one more thing to your equations though: Core loading. A BD module is about 1.75x a SB core, true, but you need to load the ENTIRE module to 100% to reach that point. As a result, Intel will perform better up until 6 of BD's 8 cores are used (3 BD modules).
Score
0
a c 79 à CPUs
January 25, 2013 4:00:16 PM

^ Totally agree. It is a well established that lightly threaded applications are really quick on Intels and heavy and highly threaded apps execute better on AMD's. The only thing favouring Intel right now is that transition to heavy multi threaded approach in software programming is happening at a slower pace and the margin that Intel has over AMD in light applications is pretty significant.More than what AMD has over Intel in (so called)wellthreaded apps.
Score
0
a b à CPUs
January 25, 2013 4:10:08 PM

^^ Which again is partially due to the fact the second core of a BD module is only about 75% the strength of a stand alone CPU core, so a 8xxx BD really only performs about as well as a native 6 core chip would. Throw on Intels superior IPC, and you see why you need to get to more then 6 cores loaded for AMD to produce a performance advantage.
Score
0
January 25, 2013 7:37:24 PM

Now, just when i though this thread was dead, real good stuff is coming though, thanks for all the input, guys, really awesome jargonathon going here. In simple terms, those of you, who need to understand, if you want to play games - like the most high end games right now buy a dual core CPU (something newer will definetly help), brand doesn't matter! But if you are looking at 64 bit running, 32 GB demanding and non-gpu based high end system for rendering and encoding at reasonable price, go for AMD.

I don't understand why everyone keep saying that Intel i7 is the best for gaming, seriously it's not, a good DX11 GPU is good for gaming, a ridiculously expensive GeForce GTX 680 is going to kill any game on planet for the next 5 years, even on a core 2 duo config.

Nobody will buy a CPU for its IGP, HD 4000, 3000, 2500, 2000 are all gimmicks, doesn't really add to the performance of the CPU, only the price, if you do, it's not worth it. Most applications don't recognize them, as yet, After Effects for instance doesn't utilize them as standard GPU, nor does the Opel GL works.

I am an Intel user, since Pentium III days, never bought anything but Intel CPU, recently i bought i5 3450. My requirement - to suck out the last bit of performance through Adobe Suite 64 Bit, 8 GB ram, and two seagate 2 TB and 3 TB HDD. Am i happy with the performance of my system? Yes, but will I be more satisfied if i had a 8 Core AMD plus a dedicated DX11 GPU combined at the price of one i5 - i will take that. The only thing that stopped me from buying an AMD CPU was high power consumption and heat. Trust me, where i live they are really the reasons that matter.

True story!
Score
0
January 28, 2013 4:05:18 PM

Best answer selected by austenwhd.
Score
0
!