Devil’s Canyon proves Intel doesn’t care about PC enthusiasts .

[LukaP]

I'll take an 8350 any day over a Devil's Canyon. You know what !? 5Ghz on air is actually COMMON for 8350s.

yes, but with performance equal or less than a 4GHz i5... so whats the point... most people need, and will for some time, need good IPC, rather than a gajillon cores... and those who need all the cores, need speed too, so they go either extreme editon. or xeon. the 83xx CPUs are a useless AMD venture that should be stopped after the first generation.

[TERAFLOP]

yes, but with performance equal or less than a 4GHz i5... so whats the point... most people need, and will for some time, need good IPC, rather than a gajillon cores... and those who need all the cores, need speed too, so they go either extreme editon. or xeon. the 83xx CPUs are a useless AMD venture that should be stopped after the first generation.

Not true, clock for clock 8350 beats i7 3770K but 8350 can clock much higher thanks to the clock resonant mesh that AMD is using.

Clock for clock 8350 matches devil's canyon in multicore but then again 8350 achieves much higher clock rates.

Dollar for dollar 8350 steals the show and beats everything from intel.

[LukaP]

Not true, clock for clock 8350 beats i7 3770K but 8350 can clock much higher thanks to the clock resonant mesh that AMD is using.

Dollar for dollar 8350 steals the show and beats everything from intel.

 

 

clock for clock, core for module, a 2600k beats an 8350... x3xx series has the IPC around the same as Nahalem.... they can clock higher due to larger process. but they would still need to be double the clock to be the same, thread to thread to haswell, and ive yet to see a stable 8.8GHz 83xx... while 4,4 is normal on haswell

[digitalnav]

As someone who has OCed more than 100 4770K's and over 50 4790K now, I can tell you that Devil's Canyon is DEFINITELY a huge improvement. 

[Misanthrope]

As someone who has OCed more than 100 4770K's and over 50 4790K now, I can tell you that Devil's Canyon is DEFINITELY a huge improvement. 

 

The article doesn't talks about a "huge improvements" but specific promises. I mean I understand what you're saying and you of all people probably had a lot more realistic expectations but calling out bad PR claims isn't a bad idea: people need to remember stuff like this or like those magical Nvidia drivers back a little while ago that turned up to be nothing spectacular.

[That Norwegian Guy]

As you scale transistor density you'll generally find it hard to maintain transistor frequency.

 

People had this same discussion with i7 2600K vs 3770K, the Average 2600K overclocked beyond the IPC improvement bought on by Ivy.

 

Not at all.

 

Comparing the i5-3570K and i5-4670K using Passmark's aggregated total results, these two chips at the exact same clock are 7.4003% apart in score. That means the IPC advantage of the Haswell chip amounts to 7.4003%.

Now, an Ivy Bridge CPU typically overclocks at least 10%~15% better. Thus, IMO, Ivy Bridge is strictly better for overclockers as they exhibit better end performance.

Sandy Bridge vs. Ivy Bridge however, the difference in IPC and overclocking headroom seem to be completely even. 6% and 5% respectively.

That makes Ivy Bridge the sweet spot for overclockers, even above Sandy Bridge, as the final performance is on average quite a bit better than Haswell and marginally better than Sandy Bridge.

[harrynowl]

Not at all.

 

Comparing the i5-3570K and i5-4670K using Passmark's aggregated total results, these two chips at the exact same clock are 7.4003% apart in score. That means the IPC advantage of the Haswell chip amounts to 7.4003%.

Now, an Ivy Bridge CPU typically overclocks at least 10%~15% better. Thus, IMO, Ivy Bridge is strictly better for overclockers as they exhibit better end performance.

Sandy Bridge vs. Ivy Bridge however, the difference in IPC and overclocking headroom seem to be completely even. 6% and 5% respectively.

That makes Ivy Bridge the sweet spot for overclockers, even above Sandy Bridge, as the final performance is on average quite a bit better than Haswell and marginally better than Sandy Bridge.

Depends what results you look at I suppose. Sandy was definitely more consistent in overclocking though as it didn't use crap TIM so temperatures weren't so much an issue. Most Ivy systems I see running in peoples signatures are around 4.4GHz, Sandy around 4.8. Haswell around 4.3.

 

Based on HWBot average OCs for 2500K and 3570K they are even. (Like a 60MHz edge to the 2500K). But Haswell has some architechtural improvements that just make it plain faster such as the 512bit thing that makes AVX2 float operations much faster.

[digitalnav]

The article doesn't talks about a "huge improvements" but specific promises. I mean I understand what you're saying and you of all people probably had a lot more realistic expectations but calling out bad PR claims isn't a bad idea: people need to remember stuff like this or like those magical Nvidia drivers back a little while ago that turned up to be nothing spectacular.

I do agree that things got out of hand but people didn't pay attention to the details. Everyone was expecting, "OMG 5 GHz on AIR!" after the Intel event but no one really pointed out that they hair LN2 air being put through the heatsinks and that those chips were likely highly binned.

All these new DC chips really mean is easier overclocking with better temps and  a .2 GHz on average higher speed compared to 4770K.

[Misanthrope]

I do agree that things got out of hand but people didn't pay attention to the details. Everyone was expecting, "OMG 5 GHz on AIR!" after the Intel event but no one really pointed out that they hair LN2 air being put through the heatsinks and that those chips were likely highly binned.

All these new DC chips really mean is easier overclocking with better temps and  a .2 GHz on average higher speed compared to 4770K.

 

Fair 'nuff

[GPUXPert]

clock for clock, core for module, a 2600k beats an 8350... x3xx series has the IPC around the same as Nahalem.... they can clock higher due to larger process. but they would still need to be double the clock to be the same, thread to thread to haswell, and ive yet to see a stable 8.8GHz 83xx... while 4,4 is normal on haswell

Clock for clock core for module 8350 not only beats the 2600K but also the i7 3820 which has significantly more cache & beats the 3770K all the while it can achieve higher clock rates than all of the aforementioned intel CPUs as well.

http://linustechtips.com/main/topic/48571-intel-amd-architectural-discussion-how-far-ahead-is-intel/

 

Compression : uses all integer threads available.

Hash-rate, same algorithm used to mine bitcoins : uses all integer threads available.

FPU VP8 / SinJulia : uses all floating point threads available.

7-Zip file compression : uses all threads available.

Video encoding : uses all threads available.

Photoshopt Cs6 : uses one thread.

Cinebench : single threaded

Ray-tracing renderer : uses all threads available.

 

 

So we're noticing a pattern here, we're all familiar with, single threaded workloads run faster on the larger intel cores, however anything that utilizes all threads available runs faster on the smaller more parallel architecture of the AMD module.

This is nothing new, the 4 extra logical threads (hyper-threading) of intel's i7 processors help it keep up with AMD's eight physical integer cores. Although the physical integer cores still maintain a performance lead, the degree of this depends on how well optimized the workload is for hyper-threading.

But what about Intel CPUs that don't support hyper-threading?

In that case 3 AMD modules become as fast as 4 intel cores as long as they are clocked 500mhz higher, although a 500mhz delta isn't really large considering that haswell i5s usually top at 4.3Ghz while the 6 core AMD processors can reach 4.8Ghz on the same cooling.

FX 6300 @ 3.5Ghz vs i5 4430 constant turbo on all 4 cores @ 3.0Ghz.

FX 6350 @ 3.9Ghz vs i5 4570 constant turbo on all 4 cores @ 3.4Ghz

[Victorious Secret]

All you Intel and AMD fanboys keep your cocks out of this thread. This isn't about which processor is better or which is weaker, and all of you children need to stop derailing every single thread with your penis meausring contests. 

 

Congrats. Your i7 is better. Or your 8320 is better. Who gives a shit in the real world. Stop derailing and polluting these threads. And any other thread.

 

I'd turn "X CPU is better than Y CPU" derailment posts as bannable offense if I could. You guys never stop till you've successfully derailed the thread into a playground full of innocent children.

[MilkyWhite]

Clock for clock core for module 8350 not only beats the 2600K but also the i7 3820 which has significantly more cache & beats the 3770K all the while it can achieve higher clock rates than all of the aforementioned intel CPUs as well.

http://linustechtips.com/main/topic/48571-intel-amd-architectural-discussion-how-far-ahead-is-intel/

 

Compression : uses all integer threads available.

Hash-rate, same algorithm used to mine bitcoins : uses all integer threads available.

FPU VP8 / SinJulia : uses all floating point threads available.

7-Zip file compression : uses all threads available.

Video encoding : uses all threads available.

Photoshopt Cs6 : uses one thread.

Cinebench : single threaded

Ray-tracing renderer : uses all threads available.

Very intriguing article, thanks for sharing !

[Phillyphries]

I don't think Intel will do anything until AMD steps back into the game somehow tbh

 

All companies do this. They sit at the top and release refreshes every year with very very minimal improvements

[IdeaStormer]

Don't like it, don't buy it. Write a letter to Intel and tell them your issues with their new tech!

 

Not going to do much unless you were a vendor selling 100 thousand or so CPU's, but you may feel better after you vent to them. Start looking at AMD.

[Faa]

Not true, clock for clock 8350 beats i7 3770K but 8350 can clock much higher thanks to the clock resonant mesh that AMD is using.

 

Wait, the 8350 is better than the 3770K?

 

Clock for clock 8350 matches devil's canyon in multicore but then again 8350 achieves much higher clock rates.

 

Lets wait again, 8350 outperforms a 4770k? 

 

Dollar for dollar 8350 steals the show and beats everything from intel.

Saying the 8350 is better than the 3770k clock for clock perfectly comes down that you claimed their IPC is better. Most people use a desktop for gaming or just general mainstream stuff, it just all boils down to singlethreaded performance except for a very minor group who do productivity. There's no such thing as multithreaded performance/price for normal desktop users. The price difference in usa isn't there atm, Intel offers a much better value. The 4670k is much faster than a 8350 for gaming, it's multithreaded performance is mostly just 5% behind AMD so it's definitely a better value even i3's are a better value.

Currently AMD's better price/performance is non-existent, people would put that virtual multithreaded performance into a price ratio unless you claim a single thread can be executed on an unlimited amount of cores at the same time you might have a point.

 

 x3xx series has the IPC around the same as Nahalem....

A E8500 at 4.7GHz outperformed a 8350 at 5.4GHz in cinebench single threaded, not sure how guy get his ideas from >.> Claiming a 8350 outperforms a 3770k clock for clock.

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=1

[BookerDeWitt]

I don't think Intel will do anything until AMD steps back into the game somehow tbh

 

All companies do this. They sit at the top and release refreshes every year with very very minimal improvements

I agree if AMD had a stronger position in the market when Haswell launched all the issues that popped up at the time ranging from Intel cherry picking golden chips to give to reviewers to the insane heat issues with Haswell and the reduced or static efficiency of Haswell under load in comparison to Ivy bridge. All of these issues would've been raised & smashed by AMD had they had a stronger market position.

 

Don't like it, don't buy it. Write a letter to Intel and tell them your issues with their new tech!

 

Not going to do much unless you were a vendor selling 100 thousand or so CPU's, but you may feel better after you vent to them. Start looking at AMD.

That's exactly right, vote with your money. If you don't like what Intel, Nvidia or AMD are doing don't buy their product, give your money to their competitor. After I learned about Nvidia's tier 0 scandal where they were basically giving money to builders to bad mouth AMD and then when they escalated with GameWorks I decided that my next upgrade should not be from them simply to send a very strong message that I don't agree with what they're doing. Now again with intel, for the longest time they were buying silicone precursors from warlords in occupied countries because it was cheaper they also built factories on stolen lands in the middle east we can't just stand by & say we don't like it Intel please stop it. What we actually do is tell them no Intel, if you continue doing this we will boycott your product.

Examples extend to Samsung bribing reviewers and covering up work related deaths in their factories, Apple hiring children to build their products, on & on & on. We must make a stand !

[TERAFLOP]

A E8500 at 4.7GHz outperformed a 8350 at 5.4GHz in cinebench single threaded, not sure how guy get his ideas from >.> Claiming a 8350 outperforms a 3770k clock for clock.

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=1

 

Cinebench #1 is a synthetic benchmark which doesn't reflect any real world useage.

#2 Cinebench uses an intel Compiler which includes a malicious code path which specifically targets non-intel CPUs to cripple performance.

http://3d-coat.com/forum/index.php?showtopic=14916

Windell talked about the Cinebench scandal & provided documentation.

[Faa]

Cinebench #1 is a synthetic benchmark which doesn't reflect any real world useage.

#2 Cinebench uses an intel Compiler which includes a malicious code path which specifically targets non-intel CPUs to cripple performance.

http://3d-coat.com/forum/index.php?showtopic=14916

Maybe you should take the effort to read the PDF file. http://www.ftc.gov/sites/default/files/documents/cases/091216intelcmpt.pdf

There isn't any proof, just as usual Teksyndicate brainwashed you. Intel's compiler even performed better on AMD cpu's than with AMD's own compiler

It doesn't reflect any real world usage. Lovely logic, tell me why there's a bunch of difference between AMD & Intel in cpu bound games? You see gaming benchmarks showing a 100% difference, Cinebench doesnt even show a 100% difference single threaded. Your 8350 is just a 2006 CPU with a bunch of bugsbunny cores, you just don't want to admit it. 8350 with 4 more cores and a higher clock speed of 1000MHz just a silly 25% faster than a QX9650 2008 cpu, clearly its IPC is lower.

[TheProfosist]

Nothing new here? I mean they do PC better but just a bit. I should also make note that that's still the mainstream line not the enthusiast one which still has the CPUs soldered. A 10% jump is about average, its been like that for quite some time. Hz isn't everything and in trying to make them more efficient they don't on as well? This doesn't surprise me in the least. K series processors were around prior to 2010 I think they go back to the C2D and C2Q revision if I recall. Intel has also said the improvements will benefit air OCers more that extreme.

[dalekphalm]

--snip--

 

Granted most enthusiasts wouldn't use FM2+, but LGA1150? What the hell? You realize that 90% of enthusiasts use LGA1150/1155, etc? That's actually the core enthusiast platform. Most can't afford, or wouldn't need LGA2011.

 

I assume you're either a troll or ignorant to the facts. If the latter, I apologize but you should definitely search this topic more before commenting.

[sof006]

Same can be said for my FX 8320, yet it's half the price.

Yet mine has a longer life span. You're more likely to upgrade again in the near future. I however will probably not need to upgrade for a long time. There are still a lot of people out there still using the first ever i7's etc that were released years ago.

[dalekphalm]

Yet mine has a longer life span. You're more likely to upgrade again in the near future. I however will probably not need to upgrade for a long time. There are still a lot of people out there still using the first ever i7's etc that were released years ago.

Agreed, I'm one of them. I'm using a Xeon w3520, which is identical to the i7-920. This beast is still more than enough for daily tasks, and I've only JUST started to consider an upgrade to Devil's Canyon specifically actually. Most of the issues with my CPU is actually the old chipset with lack of USB 3.0 and so on.

[kirashi]

Just you average "poor" but happy AMD user over here.

I won't argue the performance metrics between Intel and AMD's offerings:

We've seen time and time again Intel outperform AMD in most everyday situations.

And we've seen AMD outperform Intel with an APU in certain graphical tasks.

 

What I will argue is that you can't miss that which you've never had. Sex, drugs, rock 'n roll, and now, the CPU race. What NeXT?

[deathjester]

Wait people are mad they came out with a new chip that is slightly faster runs a bit cooler and is pretty much the same price?

 

Or are they mad about a dual core that can play games well and microcenter and newegg have been running deals with z97 AND CPU for 100 bucks?

 

It isn't Intel's fault that our API sucks and duo cores are baller until DirectX 12 comes out. That would be Microsoft's fault for taking this long to give us a low level API, and then making us wait until 2015 holiday for it, so they can sell their new OS and crappy Xbox One's until then lol.

 

I have no idea what dreamworld people are living in where all 4770k's hit 4.5-4.7 ghz but it isn't here. They could struggle to get 4.2 before they ran into heat problems and even water cooled vendors like Origin PC would only guarantee 4.1 on stupidly expensive custom builds.

 

Yup my 4770k does 4.3 at 1.15v and runs pretty cool on a evo 212. My best friends (on a much more expensive Asus board) is at 4.2 at 1.3v on water. I can do 4.5 at 1.22v. That sure isn't as hell the rule but the exception. People with the new chip basically got higher binned 4770k's that also run a little cooler. That is a hell of a lot better than a possible low bin that runs hot as hell. If anything 4770k was the screwup, but then again it crushed the competition even near stock, so how did they screw up?

 

I am REALLY cracking up at the AMD fanatics in this thread. People can now buy a 100 dollar mb/chip that runs games better or as well as the AMD chips (and will until our API isn't a joke) that can also upgrade to the next series intel chip, oh and they can run things like Dolphin Emulator like a boss, that a AMD couldn't run at 7 ghz lol. I wouldn't touch AMD now at ANY price range for gaming. 

 

BTW I wouldn't touch an Nvidia GPU until a 780ti atm. AMD all the way. Not because I am a fanboy, but because the price/performance of Nvidia atm = Nvidia is on drugs. Intel? They are killin it. Best budget build (that can be upgraded to boot), best high end build. What more do you want? 

[totjup5]

Intel will drop the same shit every year as long as AMD isn't pressuring them to do anything.