Best AMD Cpu vs Intel i7 980

[i_build_nanosuits]

Granted the Core 2 line does lack some instructions and has a slower cache (my only real issues is the lack of instructions), but as for the amount of L2, with Wolfdale and Yorkfield they had 3MB per core, while the FX line has 2MB.

but the FX have a massive shared L3 cache pool to compensate...and it's fast.

[Faa]

lol at this noob. first off that bench is for a 980x extreme and op is talking about a regular 980. 

Difference is a few more multipliers that's all, FSB can still be heavily overclocked on a 980, so other than the extra multipliers theyre exactly the same.

 

second you cant go by something as unreliable as cinebench. 

The Teksyndicate video was proven to be fabricated and paid by AMD. No retard, will ever claim AMD's singlethreaded performance is 4x faster than Intel when a freaking half million people watched that video unless you're paid to do so.

 

dont wanna watch? pretty much cinebench was designed to make intel go the easiest route and force amd to go the absolute hardest route to make intel appear much higher than it should be..there was even legal documents of it and a patch (that you manually have to install ) for it.

Why not link a video who's results are matching any other source have provided?

Right, all CPU's at 2.8GHz;

We see a 2005 architecture outperforming a 8350 at 4GHz. Now with the 8350 at stock speeds and a Q6400 clocked at 2.4GHz;

So back off with your "Intel sponsored" BS, most benchmarks out there show the 8350 doing worse than in Cinebench. 

 

cinebench will have people thinking the x5650 is the best cpu which is far from true lol

It is, two QX9650's on a dual socket board at 4.2GHz will wipe the floor with the 9590.

 

also..i said the 980 was better....but not by much which at stock is true. i also said if both overclocked the 980 wins.  not sure where you're going with all this.

 

You do realize a 980 only has 17% less clock-for-clock performance than a 3930K, right? Not by much would mean 5-10%, you must be a fool to think a 3930K is only 20-25% faster than a 8350.

http://be.hardware.info/reviews/4360/15/45-processors-review-van-intel-celeron-tot-core-i7-van-amd-a4-tot-fx-benchmarks-igpu-cinebench-115

http://be.hardware.info/reviews/4360/16/45-processors-review-van-intel-celeron-tot-core-i7-van-amd-a4-tot-fx-benchmarks-igpu-tech-arp-x264-hd-501-pass-1

http://be.hardware.info/reviews/4360/17/45-processors-review-van-intel-celeron-tot-core-i7-van-amd-a4-tot-fx-benchmarks-igpu-tech-arp-x264-hd-501-pass-2

http://be.hardware.info/reviews/4360/18/45-processors-review-van-intel-celeron-tot-core-i7-van-amd-a4-tot-fx-benchmarks-igpu-cyberlink-espresso-67

 

 

come at me bro..but you better come correct with your BS

You have been lying to people in a few older threads where people heavily criticized you for spraying misinformation to do AMD a favour. It's against the CoC, when are you going to quit trying to offer people an AMD cpu like it's your d*ck?

 

 

you guys keep bringing up single core speed..you realize thats one aspect of a cpu right?

 

It's the only aspect of a CPU that's worth discussing. You judge CPU's by their IPC along with the power consumption. Corecount doesn't matter because every core is exactly the same, so if you have 8 cores you'll have up to 8x more performance if scaling is lineair. Corecount especially for AMD is nothing more or less than a sign "we can't do shit anymore". AMD hasn't brought a 16 core out yet because it has to be near 4GHz so it would be able to play a game (2GHz FX is quite shit, 50% slower than Conroe) and the power consumption would be so bad that it wouldn't allow them to get near 4GHz meaning it won't even do much better than the 8350/9590 in multithreaded performance. And even if they manage to do it, they'll charge another 150$ for a 360mm AIO. Even the best current AMD board won't be able to overclock such a CPU that's pulling 500W from the CPU cable, let alone the mosfets staying cool enough so the CPU doesn't throttle. 

Lets judge how good the 8350 actually is against a CPU that has the same corecount;

Sorry for using a 10 years old architecture, I'm really sorry. 1.2GHz lower clock speed, architecture is 10 from 2005 and your almighty 980 killer got rekt like shit.

 

 

what about integer math, floating point , prime numbers ,extended instructions , compression, encryption , physics and sorting?

 

In none of those workloads AMD is faster. There are only two kinds of calculations; Integer/Floating point.

 

 

then how come at stock the 8350 actually has faster single core speed than the 980?

 6 core vs 8 core multithreaded, 6 core wins and you claim the 8 core (8350) has more singlethreaded performance. Sorry but even a 5 years old kid wouldn't fail at this.

 

  

The FX is great at integer instructions because it has 8 integer units (2 per modules)

 

No, 16 ALU's (4 per module). Nehalem has 3 ALU's per core, so a total of 18 versus 16; http://www.realworldtech.com/includes/images/articles/sandy-bridge-5.png?71da3d

[i_build_nanosuits]

 

No, 16 ALU's (4 per module). Nehalem has 3 ALU's per core, so a total of 18 versus 16; http://www.realworldtech.com/includes/images/articles/sandy-bridge-5.png?71da3d

 

so 2 fpu and 4 alu per modules? i dont know why i always thought it was 1 and 2...

[Faa]

so 2 fpu and 4 alu per modules? i dont know why i always thought it was 1 and 2...

1 thing; never take AMD PR serious. Just never. 4 FPU's and 4 ALU's per module yes.

It's only showing one ALU cluster (what people call a core).

AMD's pr;

Misses the MMX units..

Still doesn't say there are two 128bit MMX units >.>

[i_build_nanosuits]

1 thing; never take AMD PR serious. Just never. 4 FPU's and 4 ALU's per module yes.

It's only showing one ALU cluster (what people call a core).

AMD's pr;

Misses the MMX units..

Still doesn't say there are two 128bit MMX units >.>

even more confused now...can you sumarise that as clear as possible please? if they have the same amount of fpu as there are alu's per module then it would mean they would be true 8 core cpu's and yet they are not right?

[Faa]

even more confused now...can you sumarise that as clear as possible please? if they have the same amount of fpu as there are alu's per module then it would mean they would be true 8 core cpu's and yet they are not right?

The first X86 CPU's didnt even have an FPU, even if your 50 core CPU doesn't have any FPU it doesn't make it a fake 50-core. It will just process floatingpoints quite slow and won't have the simd (another word for FPU) extensions like AVX, AVX2, SSE etc. FPU's arent a requirement for x86 and they don't decide a CPU to be a true x-core or not. However an ALU is required.

I'll explain later why it's not a "real" 8-core.

The integer cluster what people call integer core or just a core (I call this ALU cluster) just has 2 ALU's -> 4 per module.

The FPU has;

- Two 128bit FMAC FPU's (they can work together as a single 256bit singlethreaded mode)

- Two 128 bit MMX FPU's

Thats it. Back-end is made out of things like ALU/FPU etc. Front-end is the upper side instruction decoder, dispatcher etc etc the part that tells the back-end what to do. The thing that it makes it a fake 8 core is because it's using a single front-end rather than two, not because of the FPU count. The 5960x obviously has 8 front-ends therefore making it a real 8-core, AMD's "16"-core opteron CPU's have 8 front-ends.

What AMD did was just duplicating the ALU cluster, compare the right one with the left one;

AMD just duplicated one ALU cluster therefore making it CMT and gave them a reason to call their 8350's a 8-core. You might ask now why AMD didn't use 2 dedicated front-ends to make it a real 8 core; well if they did they wouldn't have saved space. That's what CMT is about; saving space. Multithreaded performance would have been much better if AMD used 8 front-ends rather than 4.

[i_build_nanosuits]

snip!

alright that clear things up...but raise a new question:

why is there such a need to save ''space'' on the die? couldnt they just use a bigger die and call it a day?

am3+ socket could have just been twice as large like 2011 cpu's and house a 2x2 inch monster cpu?

[I/O]

Well I am considering selling it and building an AMD system for streaming/ video editing, and stuff like that. The AMD CPUs are pretty cheap so honestly I could probably sell this and not spend much at all to trade.

You would be downgrading hardcore going with AMD FX LOL.

[Faa]

why is there such a need to save ''space'' on the die? couldnt they just use a bigger die and call it a day?

Power consumption, power consumption and power consumption. Because AMD just sucks in terms of diesize/power consumption/performance. 

Lets just compare the 3930K with the 8350, actually the 980 would be a better example because it doesn't have an integrated pcie controller.

3930K; Diesize -> 435mm², corecount 8 with 2 being lasered cut (yes it's a 8-core), 3 ALU's per core and 6 FPU's per core and has an integrated PCI express controller. Total of 24 ALU's & 48 FPU's. You know what performance that thing can deliver with all 8 cores up, cut something like 20% of a 5960x.

8350; 315mm² -> 4 front-ends so corecount 4, no integrated pcie controller, 4 ALU's per module and 4 FPU's per module. Makes a total of 16 ALU's and 16 FPU's.

The 980 is actually 215mm² big, also 32nm, 3 ALU's and 6 FPU's as well and still faster than the 8350 by a good amount.

 

 

am3+ socket could have just been twice as large like 2011 cpu's and house a 2x2 inch monster cpu?

 

Can be achieved with more cores or just more execution resources. Would be like 800mm², not sure with what you're going to cool that thing though.. A 500W CPU doesn't suit a desktop pc.

If they used 8 front-ends it would have increased the power consumption (not by a lot), if they had lets say 16 ALU's per module or 12 FPU's per module, that power consumption would have been hilariously high between 3-4GHz to a point that you can't aircool it. Such a CPU I would guess would be clocked at 2GHz, probably an awful clocker too and consuming a shitload more when overclocked. Without improving Bulldozers current front-end, yes AMD's front-end is just garbage, you won't get that much of a performance boost by filling it up with more execution resources because the BD's front-end is just a plain bottleneck. Even if the front-end is perfect, a single thread can't always keep all resources busy so thats when clock speed for example can make a bigger difference.

[Dabombinable]

Power consumption, power consumption and power consumption. Because AMD just sucks in terms of diesize/power consumption/performance. 

Lets just compare the 3930K with the 8350, actually the 980 would be a better example because it doesn't have an integrated pcie controller.

3930K; Diesize -> 435mm², corecount 8 with 2 being lasered cut (yes it's a 8-core), 3 ALU's per core and 6 FPU's per core and has an integrated PCI express controller. Total of 24 ALU's & 48 FPU's. You know what performance that thing can deliver with all 8 cores up, cut something like 20% of a 5960x.

8350; 315mm² -> 4 front-ends so corecount 4, no integrated pcie controller, 4 ALU's per module and 4 FPU's per module. Makes a total of 16 ALU's and 16 FPU's.

The 980 is actually 215mm² big, also 32nm, 3 ALU's and 6 FPU's as well and still faster than the 8350 by a good amount.

 

 

 

Can be achieved with more cores or just more execution resources. Would be like 800mm², not sure with what you're going to cool that thing though.. A 500W CPU doesn't suit a desktop pc.

If they used 8 front-ends it would have increased the power consumption (not by a lot), if they had lets say 16 ALU's per module or 12 FPU's per module, that power consumption would have been hilariously high between 3-4GHz to a point that you can't aircool it. Such a CPU I would guess would be clocked at 2GHz, probably an awful clocker too and consuming a shitload more when overclocked. Without improving Bulldozers current front-end, yes AMD's front-end is just garbage, you won't get that much of a performance boost by filling it up with more execution resources because the BD's front-end is just a plain bottleneck. Even if the front-end is perfect, a single thread can't always keep all resources busy so thats when clock speed for example can make a bigger difference.

Compare an FX line CPU to a Pentium 4, and you'll find that there are a few similarities in the way AMD has implemented the architecture of the FX line (its different obviously, but the FX line has the same problems as P4s (Netburst)).

[Faa]

Compare an FX line CPU to a Pentium 4, and you'll find that there are a few similarities in the way AMD has implemented the architecture of the FX line (its different obviously, but the FX line has the same problems as P4s (Netburst)).

They both are x86 CPU's so there are always some similarities such as Intel ALU's are exactly the same as AMD's ALU's, always been like that. P4 is massively different than Bulldozer, P4 was nothing more or less than garbage and it took them 5 years quitting the GHz race. You're only looking at the hardware side, software that's built in the CPU such as very complex algorithms in the front-end can give you a huge performance advantage. Intel does this.

http://www.realworldtech.com/sandy-bridge/2/ 

Page 2-7 you'll see a lot of similarities between a FX & Sandy Bridge.

[Dabombinable]

They both are x86 CPU's so there are always some similarities such as Intel ALU's are exactly the same as AMD's ALU's, always been like that. P4 is massively different than Bulldozer, P4 was nothing more or less than garbage and it took them 5 years quitting the GHz race. You're only looking at the hardware side, software that's built in the CPU such as very complex algorithms in the front-end can give you a huge performance advantage. Intel does this.

http://www.realworldtech.com/sandy-bridge/2/ 

Page 2-7 you'll see a lot of similarities between a FX & Sandy Bridge.

I was hinting at the long pipeline, and the fact that just like an FX cpu, Pentium 4s could get massive overclocks for only a small performance boost. 404 error on that link BTW.

[Faa]

I was hinting at the long pipeline, and the fact that just like an FX cpu, Pentium 4s could get massive overclocks for only a small performance boost. 404 error on that link BTW.

realworldtech.com/sandy-bridge/2/

[D3MON]

the £990 cpu is still faster, who'd have thought :rolleyes: 
id keep it if i had it.

edit-£500

[Faa]

the £990 cpu is still faster, who'd have thought :rolleyes: 

id keep it if i had it.

It was 500-600$, basically you were paying another 500$ for more multipliers. What a joke but isn't the first one, Q series the mainstream ones were around 200$, QX series with more multipliers were 1000$.

[D3MON]

It was 500-600$, basically you were paying another 500$ for more multipliers. What a joke but isn't the first one, Q series the mainstream ones were around 200$, QX series with more multipliers were 1000$.

yeah i tried to find it but couldnt find uk prices, id probably of snapped that up if it was back then (same with 5820k now), i have no problem bclk overclocking.

[Faa]

yeah i tried to find it but couldnt find uk prices, id probably of snapped that up if it was back then (same with 5820k now), i have no problem bclk overclocking.

Bus clocks much higher as well though. Hated bus clocking, if your ram couldn't take 50 more MHz when your CPU was stable at 4GHz at 1.20V it really sucked. When are you getting a 5820K? 

[patrickjp93]

so 2 fpu and 4 alu per modules? i dont know why i always thought it was 1 and 2...

It's a quirky 1 and 4. Since the FPU has to check in with the other core to ensure no 256-bit wide vector instruction is incoming before reserving a smaller job, the float operations on the Bulldozer derivatives are agonizingly slow. Even adds take 8 cycles. Multiply and divide take 11.

[I/O]

Even the best AMD CPU are purly just mid range performance and price. Sometimes the performance is not even mid range but the price is always the same.

[WaitForMeImOnDialUp]

WaitForMeImOnDialUp, on 10 Jan 2015 - 10:33 AM, said:

come at me bro..but you better come correct with your BS

*Faa- You have been lying to people in a few older threads where people heavily criticized you for spraying misinformation to do AMD a favour. It's against the CoC, when are you going to quit trying to offer people an AMD cpu like it's your d*ck?

 

 

isn't that all you're gonna do FAA...daydream about my d*ck all day? and by people..you mean you,fate,and facepalm otherwise known as LTT's amd nazis

 

i NEVER lied about anything with amd. i said they are good for gaming..they are. you are the one lying..saying they are shit. 

 

i said they wouldnt bottleneck two 970's...in some games/apps it doesnt..in most it will,i'll admit it.

 

at stock speed the 8350 has higher single core speed than the 980...that is fact per passmark single score tests which dabobminable even provided.yes when overclocked it beats it..i already said that 3 times

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 11:16 AM, said:

what about integer math, floating point , prime numbers ,extended instructions , compression, encryption , physics and sorting?

 

*Faa - In none of those workloads AMD is faster. There are only two kinds of calculations; Integer/Floating point.

 

 

 a 8350 oc'd vs stock 3770k the 8350 beats it in half those things listed. only two calculations? well oc'd 8350 beats stock 3770k  at integar and come very close to matching it in floating point. clock for clock 3770k dominates, buts its twice the cost of an fx 8350.

 

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 10:33 AM, said:

cinebench will have people thinking the x5650 is the best cpu which is far from true lol

*Faa- It is, two QX9650's on a dual socket board at 4.2GHz will wipe the floor with the 9590.

 

uhmmm...thats still not the best cpu..unless you're trying to imply the x5650 or two qx9650's beats haswell-e lol, if that were the case than everyone would buy dual 771 motherboards and x5450's. but its not..and once again you are wrong.

 

also, there is no dual socket motherboard for 2 qx9650's, thats like me saying 2 fx 8350's on a dual socket motherboard beats 4790k. :rolleyes: you cant come at me with fairytale pc parts that dont exist bro.

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 10:33 AM, said:

lol at this noob. first off that bench is for a 980x extreme and op is talking about a regular 980. 

*Faa- Difference is a few more multipliers that's all, FSB can still be heavily overclocked on a 980, so other than the extra multipliers theyre exactly the same.

 

so you were wrong again? cool.

 

this is too easy

 

[Dabombinable]

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 10:33 AM, said:

*Faa- You have been lying to people in a few older threads where people heavily criticized you for spraying misinformation to do AMD a favour. It's against the CoC, when are you going to quit trying to offer people an AMD cpu like it's your d*ck?

 

 

isn't that all you're gonna do FAA...daydream about my d*ck all day?

 

i NEVER lied about anything with amd. i said they are good for gaming..they are. you are the one lying..saying they are shit. 

 

i said they wouldnt bottleneck two 970's...in some games/apps it doesnt..in most it will,i'll admit it.

 

at stock speed the 8350 has higher single core speed than the 980...that is fact per passmark single score tests which dabobminable even provided.yes when overclocked it beats it..i already said that 3 times

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 11:16 AM, said:

 

*Faa - In none of those workloads AMD is faster. There are only two kinds of calculations; Integer/Floating point.

 

 

 a 8350 oc'd vs stock 3770k the 8350 beats it in half those things listed. only two calculations? well oc'd 8350 beats stock 3770k  at integar and come very close to matching it in floating point. clock for clock 3770k dominates, buts its twice the cost of an fx 8350.

 

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 10:33 AM, said:

*Faa- It is, two QX9650's on a dual socket board at 4.2GHz will wipe the floor with the 9590.

 

uhmmm...thats still not the best cpu..unless you're trying to imply two qx9650's beats haswell-e lol, if that were the case than everyone would buy dual 771 motherboards and x5450's. but its not..and once again you are wrong.also, there is no dual socket motherboard for 2 qx9650's, thats like me saying 2 fx 8350's on a dual socket motherboard beats 4790k. :rolleyes: you cant come at me with fairytale pc parts that dont exist bro.

 

WaitForMeImOnDialUp, on 10 Jan 2015 - 10:33 AM, said:

*Faa- Difference is a few more multipliers that's all, FSB can still be heavily overclocked on a 980, so other than the extra multipliers theyre exactly the same.

 

so you were wrong again? cool.

 

this is too easy

 

 

The only reason FX CPUS manage to match any of the current i5-i7 line is purely through  a massive overclock. And that creates shitloads of heat and consumes massive amounts of power which requires a liquid cooler. And your talking about an overclocked FX 8350 beating an Intel CPU at stock, as soon as the said Intel CPU is at the same frequency, it will still kick the shit out of an FX, and the 3770k already starts out at a lower clock speed. Another thing is you denying the simple fact that at the same clock speed as any FX CPU, a core 2 duo has a higher single core speed. Re-posting the evidence:

[WaitForMeImOnDialUp]

The only reason FX CPUS manage to match any of the current i5-i7 line is purely through  a massive overclock. And that creates shitloads of heat and consumes massive amounts of power which requires a liquid cooler. And your talking about an overclocked FX 8350 beating an Intel CPU at stock, as soon as the said Intel CPU is at the same frequency, it will still kick the shit out of an FX, and the 3770k already starts out at a lower clock speed. Another thing is you denying the simple fact that at the same clock speed as any FX CPU, a core 2 duo has a higher single core speed. Re-posting the evidence:

1.JPG

2.JPG

3.JPG

4.JPG

5.JPG

6.JPG

 

yes the e8400 is one of the best overclockers in history of cpu. it was a great chip thats for sure. but that chip..maxed freq still bottlecks any high end gpu in lots of things. especially next gen multiplayer games,but yes it does have faster single core speed, it has faster single core speed than most  intel cpus,so saying amd is crap because its lesser than would mean lots of i5's are crap too...the e8400 is a damn good chip

 

but its still a lga 775 chip, which most motherboards are only ddr2 (some are ddr3) and pcie is 1.1.

 

yes im comparing a overclocked fx being very close to a stock 3770k, because for less than half the price it shouldnt even compare even if overclocked on ln2..but it does. of course 3770k overclocked beats it..its a 330 dollar cpu if it didnt then that'd be terrible

 

for good performance on a fx 8 core you do have to overclock the crap out of it..but its all good because amd fx are made to be extreme overclockers...so good that the 8350 has the world record overclock at 8.7ghz. (on liquid nitrogen so that point is meaningless but still)

[Dabombinable]

yes the e8400 is one of the best overclockers in history of cpu. it was a great chip thats for sure. but that chip..maxed freq still bottlecks any high end gpu in lots of things. especially next gen multiplayer games,but yes it does have faster single core speed, it has faster single core speed than most  intel cpus,so saying amd is crap because its lesser than would mean lots of i5's are crap too...the e8400 is a damn good chip

 

yes im comparing a overclocked fx being very close to a stock 3770k, because for less than half the price it shouldnt even compare even if overclocked on ln2..but it does. of course 3770k overclocked beats it..its a 330 dollar cpu if it didnt then that'd be terrible

 

for good performance on a fx 8 core you do have to overclock the crap out of it..but its all good because amd fx are made to be extreme overclockers...so good that the 8350 has the world record overclock at 8.7ghz. (on liquid nitrogen so that point is meaningless but still)

That's only 700MHz more than the Pentium 4s record.

[D3MON]

Bus clocks much higher as well though. Hated bus clocking, if your ram couldn't take 50 more MHz when your CPU was stable at 4GHz at 1.20V it really sucked. When are you getting a 5820K? 

when money pours out my a55 or i get bored enough, ive got nothing else to upgrade now and I'm a benching whore..

going from what i can gather from cpu-world, 133fsbx25=3.33ghz with 1333 ram divider

160x25=4ghzwith 1600 ram and 1333 divider.

166x25=4.15ghz with 1333 ram and 1066 divider.

no ram overclocking needed, disable turbo and it should be easy..

if the 980x can reach it a 980 should unless intel really binned thier chips. i dont think amd do, from what ive seen most amd chips all manage the same (unless the owners a nitwit and using a rubbish £30 board or piling in 1.6v).

[Faa]

isn't that all you're gonna do FAA...daydream about my d*ck all day? and by people..you mean you,fate,and facepalm otherwise known as LTT's amd nazis

The fact that you haven't managed to sell anyone a 8350 here yet proves that nobody was interested in your d*ck. Bet your willy called "8350" has been mostly useful when watching the Teksyndicate 8350 video. Just hail Logan the AMD prophet, fock yeah? 

 

i NEVER lied about anything with amd. i said they are good for gaming..they are. you are the one lying..saying they are shit. 

You did, plenty of evidence has been thrown before your first post and you said a 980 is only a tiny bit faster than a 8350. Also the fact that you're trying to trick people with Teksyndicate video's is also a lie. Can't be the truth, so has to be a lie.

 

i said they wouldnt bottleneck two 970's...in some games/apps it doesnt..in most it will,i'll admit it.

 

Uhm, you never said this in this thread. 

 

at stock speed the 8350 has higher single core speed than the 980...that is fact per passmark single score tests which dabobminable even provided.yes when overclocked it beats it..i already said that 3 times

O shut it. It doesn't. Reading now, they were with some massive suicide overclocks vs a stock CPU - how relevant.

 

 a 8350 oc'd vs stock 3770k the 8350 beats it in half those things listed. only two calculations? well oc'd 8350 beats stock 3770k  at integar and come very close to matching it in floating point. clock for clock 3770k dominates, buts its twice the cost of an fx 8350.

 

It's about the 980 vs 8350, not 3770K. Starting another argument that's not even ontopic.

Also 200 vs 300$ isn't twice as much, is it? 

 

uhmmm...thats still not the best cpu..unless you're trying to imply the x5650 or two qx9650's beats haswell-e lol, if that were the case than everyone would buy dual 771 motherboards and x5450's. but its not..and once again you are wrong.

I'm not saying it's the best CPU. Also a 5820K at stock outperforms two QX9650's at 4.2GHz.

 

also, there is no dual socket motherboard for 2 qx9650's, thats like me saying 2 fx 8350's on a dual socket motherboard beats 4790k.  :rolleyes: you cant come at me with fairytale pc parts that dont exist bro.

We both know I meant the QX9775, another useless argument. QX9650's are exactly the same as QX9775's except for stock clock speeds.

 

so you were wrong again? cool.

Having troubles making sense again? You haven't proved a damn thing yet.

 

this is too easy

Yeah making things up is quite easy.