Intel's Core i5 6500 Shines As a $199 Skylake Processor, Works With Linux (phoronix.com)

An anonymous reader writes: Intel has begun releasing more "Skylake" processors that are cheaper than the launch SKUs of the i5-6600K and i7-6700K. One of the new processors that is now widely available is the Core i5 6500 and it costs just $199 USD — that puts it just a few dollars more than the AMD FX-8370 and significantly less than the higher-end Skylake and Haswell CPUs. At least with Ubuntu Linux, the Core i5 6500 is showing competitive performance that for some workloads puts it faster than Core i7 Haswell/Broadwell processors and much faster than any AMD processors. The Intel Skylake CPUs are fully supported under Linux but the caveat is needing the very latest kernel otherwise there's no graphics acceleration or sound support.

Comments

Re:Penguins are for Cows

[MobileTatsu-NJG]

Can I just take a moment to tell you that I like you waaaay better than APK? Your posts are fun!

Meh.

[Moof123]

Moore's law is dead. Core count is stagnant. All recent gains seem to be in incremental power savings.

In this particular case the 6500 is a fixed 3.2 GHz, while the 6600K is 3.5 GHz as shipped. The 6600K can readily be pushed to 4.2 GHz on air, which is partly why you pay the extra K tax to be able to pull those shenanigans, and you get to leave the cheaper part in th edust.

Re:Meh.

[__aaclcg7560]

In this particular case the 6500 is a fixed 3.2 GHz

IIRC, the 6502 in my Commodore 64 was fixed at 1MHz. Nice to see old tech getting a much need speed boost! :P

Re:Meh.

[Anonymous Coward]

Guess what? Until AMD steps up and delivers a killer product at a killer price you will see nothing but stagnation from Intel.

Re:Meh.

[FlyHelicopters]

Moore's law is dead. Core count is stagnant. All recent gains seem to be in incremental power savings.

You discard the improvements in power savings like they are nothing.

Today you can get the same performance as 5 years ago, for 1/3 the power consumption.

That is a massive improvement. Moore's Law isn't dead, instead of more performance, Intel has focused on using less power to provide the same, or slightly better performance. Give these new chips 130w to play with and they'll blow away the older stuff. But what took 130w 5 years ago now only takes 45w.

That is a big deal.

Re:Meh.

[Junta]

Yeah, that's the interesting thing. If you just follow the desktop parts, performance has been modestly improving and power consumption is dipping. Meanwhile on the server side, power consumption has been more steady and the core count has been going drastically up. Sure workloads that don't scale to many cores aren't getting that much of a boost (and that is the state of most desktop platforms), but areas with multithreaded workload and/or consolidated bunch of single threaded workload have continued to benefit from advances.

Re:Meh.

[Anonymous Coward]

The Commodore 64 used a 6510. It was derived from the 6502 design, but had a few hardware feature (tri-state buffer) that the 6502 didn't have.

https://en.wikipedia.org/wiki/MOS_Technology_6510

Re:Meh.

[alvinrod]

Moore's law is about transistor density, not the number of cores or their performance.

We've reached a point where the amount of computational power in the chips that are being made is in excess of what the majority of consumers need so instead of further increasing the number of cores, which is pointless for most workloads or making the individual cores more powerful, the density improvements are being used to make smaller chips that are more power efficient because the most pervasive computers these days are those which are battery powered.

My phone has more computational power than my desktop computer from 10 years ago. Moore's law might be slowing down due to physical limitations as we move closer to chip designs with features that are only a few atoms thick, but we're still moving forward and should for a few more generations of chips even if it's not quite at the original pace, which isn't that big of a deal for most users anyways.

Re:Meh.

[oh_my_080980980]

Sad but true. You're paying for some pretty pathetic performance.

Re: Meh.

[Moof123]

Power in a server farm matters, as it does for a laptop, but for a home PC it is not nearly my top concern. Crappy integrated graphics are important for a laptop, but. It for a decent gaming machine. So the top two improvements are of near zero value to me who want a good gaming machine and are frustrated by the focus on GPU and power. I want more core and clock speed without the SEVERE increase in price Intel demands for aa 6 or 8 core CPU.

Re: Meh.

[FlyHelicopters]

I understand your point and what you want.

I'm simply pointing out that gains have been made, if not in the area you personally care about.

Re:Meh.

[FlyHelicopters]

Laptops are where the really big difference has been, IMHO...

5+ years ago, the idea of having a thin and light notebook that got 6+ hours of battery life while being useful was a fantasy.

Today, you can get a really useful laptop for a really reasonable price that has really nice power life.

The reasonable performance you can get in 15 watts today vs. 5, 10, or 15 years ago is astounding...

Re:Meh.

[pushing-robot]

Was just about to post this. The biggest gains in computer performance over the past few decades came from increasing clock speeds and eliminating delays. We're pretty much to the limit of silicon clock speeds, and delays are low enough that it's hard to find new ways to reduce them—throwing more transistors at the problem won't help. Moore's Law hasn't stopped, it's just not relevant to CPU performance anymore.

And you know what? I don't mind. Software developers only optimize until their code is 'fast enough', and the 90s and 2000s were proof of that. Computers increased in speed a thousandfold yet ran as slow and buggy as ever (and twice as hot and noisy).
Since performance began to plateau in the Core 2 days it's been a renaissance. My computers now 'feel' an order of magnitude faster than anything I had last decade, yet are incredibly quiet, slim and light, even pocketable. Devices I've owned for years are still (with the help of a SSD) perfectly usable. I'm no longer forced to upgrade every year or two just to be able to perform the same tasks at the same speed, and it feels great.

Of course, there are still situations where more performance would be beneficial, but luckily the tasks that really *need* more speed are usually very parallelizable and still benefit from more transistors. Specialized hardware like servers and GPUs are still improving at an impressive rate, and should stay on that trajectory for years to come.

Re:Meh.

[fnj]

Moore's law is dead. Core count is stagnant. All recent gains seem to be in incremental power savings.

Moore's Law says nothing about core count. It says nothing about performance. It only speaks of transistor count. How you put those added transistors to use is subject to certain realities. There are real limits to how effectively you can utilize enormous core counts in general purpose CPUs in desktop or mobile use. In servers you have enormously scaled parallel use, and can make better use of them.

It looks to me like the increased transistor count has been directed mostly at ballooning RAM cache sizes.

Improved performance/power quotient is a completely orthogonal axis from Moore's Law. It is not even implied by Moore's Law.

Re:Meh.

[pushing-robot]

Huh, I could have sworn Intel released a 4/4.4GHz chip a year and a half ago.
Or is this some copypasta I haven't seen before?

Re:Meh.

[Blaskowicz]

You are fairly wrong here, Intel has had the 4790K which is 4GHz base and 4.4GHz turbo clock, it is slightly more power hungry than the 3.x ones. 6700K is more of the same.
Or get a Pentium 3258 on H81 motherboard and set the clock at 4.4GHz, use its integrated video on linux. It's a dog bone they threw at us.

Re: Meh.

[amiga3D]

I bought a dual I-7 2ghz quad core equipped mac mini 3 years ago. In all reality it's fast enough. It's the first computer I ever felt that way about. It does everything so quick that the only time I have to wait more than a second or two is when I transcode video. That is time consuming but still pretty good. I don't game so the HD 4K graphics is adequate. After 3 years I feel no pressure to upgrade and always in the past a year after buying or building a computer I'd start craving more speed. If I was a gamer it would of course be different but my favorite game is still Railroad Tycoon 2. I think now it's just cheaper and less power that is where the most improvement comes from. For 90 percent of users out there a low end unit is more than powerful enough, even to run windows.

Re: Meh.

[laffer1]

You don't want to upgrade. The new Mac Mini is slower than the quad core. They have no replacement for the mac mini 2012 quad core model yet without spending $2000 on an iMac. I just went through this and couldn't find a single mac that was faster for under $1800.

That said, many of the new macs have SSD and better graphics but CPU performance is a joke. I decided to put a SSD into my mini instead and now I have most of the performance of a new mini in terms of IO but much faster processor.

Re: Meh.

[amiga3D]

I got a deal on a couple of 1TB 2.5 drives that rotate at 7200 RPM. I was going to buy an ssd but the mini's original drives weren't that bad so I picked up two new ones for 110 bucks and it's pretty sweet. I've got my video files on an external 4TB drive in a FW800 case so it's pretty sweet. I'll probably stick with it for another 2 or 3 years. I like the look of the new Mac Pro but it's just too pricey for a hobby.

Re:Meh.

[Hadlock]

It's nearly impossible to find a Chromebook with less than 6 hours battery life now. I took pause on buying a $150 chromebook because it "only" had 8 hours battery life, while the competing $149 laptop had 10.5 hours. The new Dell XPS 15 laptop boasts 18 hours continuous use and has a power-sucking 4K display. And yet it's about half the volume of my 2001-era Powerbook G4 that was capable of 5 hours on the lowest brightness setting.

Re:Meh.

[dave420]

Umm...

cat /proc/cpuinfo
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]
model name : Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
[...]

Re:Meh.

[Tough Love]

Moore's law is about transistor density, not the number of cores or their performance.

Actually, it's about transistor count.

Re:Meh.

[snsh]

The benchmarks for the trusty old i5-2500k are still 60-70% of the i5-6500. So from early 2011 to late 2015, CPU performance has gone up by a whopping 50%.

Moore's law is about transistor count, not performance, but even by that's only 10% higher in Skylake compared to Sandy Bridge.

Re:Meh.

[ToddInSF]

Saving power saves heat and potentially extends lifespan. So it's a big deal; arguably as important or possibly MORE important than faster cycles.

The ability to shrink size and still increase processing power while also reducing power consumption and heat generation is massive progress.

I'm wandering if anybody is arguing Moore's Law is off now because it's been exceeded.

OpenBSD?

[Anne Thwacks]

Will it work with OpenBSD?

Re:OpenBSD?

[erapert]

No. It needs a blob which includes special NSA software to run, which is only out for Linux, Windows, and MacOS.

Still waiting for prices to drop...

[__aaclcg7560]

I want a $50 processor and a $50 motherboard, which I can easily do with the 1150 chipset. For the 1151 chipset, it's $300+. Of course, that doesn't include memory. Motherboards that uses a newer processor with DDR3 memory, or an older processor with DDR4 memory, cost $150. I might have to go with a transitional motherboard in the near future to bridge price gap.

Re:Still waiting for prices to drop...

[sexconker]

Or do like me: Stop buying shit that represents only marginal improvements.
I'm sitting pretty on my 4.5 GHz i7 from 4 years ago. The only thing I feel like upgrading is my GPU.

Can we please get back to upping clock speed?

Re:Still waiting for prices to drop...

[__aaclcg7560]

The 4.5GHz is probably as good as its going to get as far as clock speed. If you haven't noticed, processors have multiple cores because the 5GHz barrier can't be broken at an economical price point. In fact, server processors are clocked at a lower clock speed to reduce power consumption and run far more multiple cores than consumer counterpart.

https://www.youtube.com/watch?v=48A_Yqj965c

Re:Still waiting for prices to drop...

[oh_my_080980980]

AMD is coming out with a new processor architecture next year. So yes Virginia it is happening. Problem is Intel's capex spending is down and the market is favoring mobile devices, so the PC market is bit of an after thought.

Re:Still waiting for prices to drop...

[jonwil]

The only reason I want to upgrade this Core 2 Duo is because Visual Studio takes forever to compile/link large projects.

Put an SSD and a new GPU in it a while back and its been going great on gaming ever since (hopefully it will be good enough to handle Fallout 4 when that hits)

Re:Still waiting for prices to drop...

[__aaclcg7560]

The last time I rebuilt my gaming PC was for Windows Vista in 2007. Since then I've added a SSD, changed out the CPU twice, and went through three different video cards. I'm planning to switch out the AMD 690 (DDR2 RAM) motherboard for a AMD 990 (DDR3 RAM) motherboard to get a few more years before I rebuild again.

Re:Ubuntu

[CrashNBrn]

So you installed an Intel Core I5 6500 into the laptop then?

Re:Penguins are for Cows

[willworkforbeer]

Ooh....get a room you two...

Or a milking stool.

Re:maria

[truck_soccer]

Maria, you have fat greasy hands.

I wish they had some reference power testing

[avandesande]

I have a I7 950 that is pretty old. It is still more than fast enough for my needs but would consider replacing it for something that uses less power. I would be curious to know how much it uses under the same testing conditions.

BTW the i5 6500 avg 40 watts.... nice!

Re:I wish they had some reference power testing

[slaker]

That old i7 is in benchmark terms competitive with current i3 CPUs. I don't think of current i3s as slow and I don't think of five year old i7s as slow either. I upgraded to a 5960k last December because I actually do enough video encoding to keep it fed, but if I'm not stealing Blu-Rays there's no subjective difference from that monster to a the i7-980 it replaced.

Re:I wish they had some reference power testing

[oh_my_080980980]

Your stating that a dual core i3 processor out performs a quad core i7 processor? How fucking slow is the i7?

Re:I wish they had some reference power testing

[FlyHelicopters]

It depends on what you're doing...

I have a i7-920 on the test bench, it is still plenty fast for most anything you'd do with your average desktop computer, but the newer chips are indeed faster.

For example, comparing the i7-920 against the i3-6320, you'll find the 2.66 GHz i7 actually slower in some tasks than the modern i3 chip.

The i3 runs at 3.9 GHz, this is more than 45% faster, and it doesn't even including the IPC improvements across that many generations.

Now, on some very specific tasks, the old i7 might be faster thanks to its triple channel memory and its 4 true cores and 8 threads.

But those situations are very specific. The dual core, quad thread i3 is enough for a lot of things and even for those where it isn't, the faster clock speed combined with the higher IPC makes up a lot of the difference.

Re:I wish they had some reference power testing

[threephaseboy]

Pretty slow.

Re:I wish they had some reference power testing

[avandesande]

The point was that I don't care about the performance at this point- just power usage. I think there are many people in the same boat as me that would upgrade their system just to save power.

Seems like a slight nudge

[truck_soccer]

If you're on haswell, don't bother. Unless you're doing it to increase the girth and length of your e-peen.

Re:Ubuntu

[Windowser]

Replacing a bloated OS with a slim one makes it appear that the hardware run faster.

Much faster than AMD: huh???

[serviscope_minor]

It's much faster in a few tests, a little faster in quite a lot more, slower in some and much slower in others. How is that "much faster"?

Re:Much faster than AMD: huh???

[gweihir]

You must not have gotten Intel's payment from the "honesty in journalism" fund. After you have, of course everything Intel is much faster!

Re:Much faster than AMD: huh???

[serviscope_minor]

Christ alive! I got modded troll! Can you believe that?

Well, due to your (entirely judtified, I might add) paranoia about intel, probably not.

It's there in the graphs. Yeah, the AMD loses some badly, but so does the intel one.

PS I'm enjoying my new FX9590s :)

No!

[s.petry]

There was no claim that the hardware ran faster, the clam was that the compute ran faster. Light weight processes, improved memory management, and less dependency on hard drive cache and paging memory and more use of real memory all make Linux run faster than Windows on the same hardware.

Even with the bloatware that is Gnome Linux is faster, and does more.

This has been a consistent benchmark for at least 15 years when common applications for CAE started running in Linux. CAD was pushed into directx so lost market there, but originally under OpenGL CAD ran better in Linux too. Would you care to guess which OS is faster for Database work, web services, etc.. etc.. on identical hardware? Be careful with what you are given in some places, because people are paid to skew the benchmarks in someone's favor. I do my own testing on my own hardware.

Re:No!

[armanox]

Would you care to guess which OS is faster for Database work, web services, etc.. etc.. on identical hardware?

Depends on the Database. I'm hearing Solaris on SPARC is supposed to be the best for Oracle DB....

Re:No!

[Trongy]

SPARC T series or M series?

Can you explain why Oracle's Exadata systems run Oracle DB on Linux and Intel x86 processors if Solaris on SPARC is so much better?

Re:No!

[armanox]

If the ODA X5-2 is any indication it has nothing to do with performance and everything to do with what the user's expect. Our M3000 boxes are more reliable and perform on par with the Exadata servers (and that says nothing of how a T4 machine does!)

Re:It would be a news story if it DIDN'T work

[beelsebob]

The compatibility question is not over the CPU, but over the GPU. While Intel doesn't really market it well, this chip comes with an integrated GPU that's faster than most of AMD's APU GPUs, while the CPU is about twice as fast as any of AMD's APUs.

Re:Ubuntu

[Anonymous Coward]

Welcome to 2010.

Who is buying the FX-8370?

[hairyfeet]

That hasn't been the sweet spot for the FX chips in like...well ever.

Us that buy AMD chips are not buying the FX-8370, the bang for the buck just isn't there, if you want an octocore the sweet spot is the FX-8320, either the FX-8320 or the FX-8320E if you want to lower the power a tad, which is what I personally went for. The really hot spot right now is the FX-6300 which is a great gaming chip. A 3.5Ghz/4.1Ghz turbo chip with 6 cores for $99? Its a kick ass buy at that price.

As for the 6500? Its a good chip, although I don't know why they are calling them "Linux friendly" when its Intel that have been coming up with all the nasty DRM chips going all the way back to HDCP and Palladium, but strictly based on the CPU? Its a good chip. I just have to wonder how many chips both Intel and AMD are gonna move when a 7 year old C2Q or Phenom II has no problem running pretty much every piece of software out there including games, both companies have built such badass chips for years that most users have piles of cycles left over.

Re:Who is buying the FX-8370?

[Lisandro]

That was my choice as well. The FX-8320E is a great CPU at a great price - Intel offerings still kill it in performance-per-core, but it is awesome for heavily multithreaded applications. You can build a system with a MSI board and 16GB RAM for under $350.

Re:Who is buying the FX-8370?

[CrashNBrn]

The Intel Core i5 6500 (no HyperThreading) appears to be faster than AMD's FX-6300 in single-process benchmarks by 40-100%, yet the multi-process benchmarks don't show as much of a pronounced difference: 10-40%. Granted the scores/graphs are all unmarked, which limits their usefulness in trying to make an informed decision, when you don't know the actual time-frame, nor if the FX-6300 was overclocked (which it can be, and the i5 can't). As you stated, TFA compares it to a similarly priced AMD processor, as opposed to the FX-6300.

I upgraded to an AMD FX-6300 6-Core Processor Black Edition ($84.99) last year to replace a Phenom II X2 560 ($99.99, from 2011). I doubt it was actually necessary, but the FX-6300 was (and is still?) the sweet spot for price/performance.

Re:Who is buying the FX-8370?

[CrashNBrn]

The FX-6300 was also a drop-in replacement for the Phenom II, on an AM3+ motherboard, which made it that much easier to justify.

Re:Who is buying the FX-8370?

[hairyfeet]

Just remember the cardinal rule which is if you don't know which compiler they used the test is worthless as Intel to this day uses the Cripple code in their compiler and pays benchmark companies to use it with predictable results. You let ME write the conditions on the compiler I can make a Sempron from 03 beat the latest i7, don't make the test worth a fuck.

If you want to see how the chip REALLY stacks up without cripple code? Just look at these Linux benchmarks and surprise! Remove the cripple code and you have The FX8 competing with i5 and i7 and the A10 beating i5...wow, isn't it amazing how much just not rigging the fucking benchmark does for the scores? Its a fricking miracle.

Re:Who is buying the FX-8370?

[the_humeister]

It's Linux friendly because the GPU drivers are open source and don't suck, unlike the AMD GPU drivers. Although if you don't do any 3D related work, any of the open source drivers for all the GPUs are good enough.

Re:Who is buying the FX-8370?

[hairyfeet]

Look at the link I provided, most benchmarks are compiled with the Intel cripple compiler which makes their results worthless. If its compiled with ICC it will scan for CPUID and if it doesn't return an Intel CPUID? It will send ALL math code through the X87 codepath, a path that hasn't been mainstream since 1995. No SSE, no AVX, none of the modern math processing is touched. How big of a difference does that make? Well if you take a Via CPU (the only one that allows softmodding of CPUID) and change its CPUID to an Intel one? BAM, you instantly gain 30% in the benches!

The FX-6300 is a really solid chip and if you scored it for $85? Man you got a hell of a deal. I went with the FX-8320E because I also do a lot of A/V work and those extra threads really helps with transcoding but for a general purpose CPU? The FX-6300 really rocks.

Oh and if you need a kick ass HTPC or second PC for dirt cheap? I've been getting 75%+ unlocks on the Athlon X3 455 which can be had for as little as $30, hellacheap for a 3.3Ghz quad, and for HTPCs you really cannot beat the socket AM1 chips, we're talking a quad core APU capable of doing 1080P for sub $80 WITH the board, great chips.

Re: Who is buying the FX-8370?

[the_humeister]

Depends on program actually and what version of ICC. What you're describing was old ICC behavior before they were sued. My own tests compiling Povray with ICC 13 I've found that the ICC dispatch code does work properly on my FX 8350. In other words, compile with AVX dispatch support and the program will choose the AVX code path, which gives the same performance as compiling with only AVX support (the dispatched program also runs on my PhenomII 1090T, but the AVX exclusive program doesn't). But that doesn't necessarily mean the ICC code is always faster. Sometimes GCC code is faster because GCC can explicitly order instructions for Bulldozer, but it doesn't do autodispatch. That has to be coded separatley by the programmer.

Re:Well

[beelsebob]

Hmm? From the charts it seems like a choice:
1) Compare against the FX8370, in which case the intel is between 5% and 100% faster; and comes with a very good GPU.
2) Compare against the A10 APU, in which case the GPU is roughly on a par, and the CPU is about 100% faster across the board.

That seems like "much faster than any AMD processors" is justified.

Lets hope for AMD Zen in Q3 2016

[future assassin]

as this price will be brutal if Intel gets the market to itself. This could be Intel trying to get a lower price going for when Zen comes out, if they get a proper price point they could hold off Zen in price vs power.

Re:Lets hope for AMD Zen in Q3 2016

[future assassin]

Well will see I love my A10 chip systems and if ZEN get come close to Intel for a cheap price I'm looking forward to new price wars on both sides maybe even push them to move forward a bit faster.

Re:Lets hope for AMD Zen in Q3 2016

[edxwelch]

The latest rumour puts Zen release date in Q4 2016.
Between then and now there are no new CPU products planned, so they have to survive on Kavari / Carrizo rebadges til then.

Re:Lets hope for AMD Zen in Q3 2016

[future assassin]

Yah just saw that today :( Gives Intel some time to get the pricing lower on some chips.

Why would anyone buy AMD?

[rsilvergun]

Not trolling. I just have a really hard time justifying the purchase. I'm running an old Athlon 6000 X2 and it rocks, but my friends with 8350s either had stability problems ( no overclocking, Asus or Gigabyte board) or just plain couldn't run their games. What I really want to know is why aren't AMD processor prices in free fall? Who's buying them that keeps the price near an i5?

Re:Why would anyone buy AMD?

[Lisandro]

Weird. I've put together several boxes using the FX-8320E and 8350 and had zero stability issues.

Pricing is a concern though. The 8320 line used to be an excellent option in the bang for the buck department, but these new Intel offerings really close the gap.

Re:Why would anyone buy AMD?

[gweihir]

I have both a gaming-box and a long-running Linux server on the FX8350. Absolutely no problems.

Re:Well

[Tsolias]

The benchmarks in the linux section show both cpus on par, except the scientific section where I can speculate that the non-existent avx2 instruction make that much difference when comparing with a 3 yo chip (but prolly it's using avx512 which _expands_ the classic AVX ISA doubling the width of the instruction from 256 to 512).
I don't see any gpu comparisons, therefore I cannot agree to your second assumptions\.
Thus, I am going to stand on my first post.

Intel supports the TPP

[flarflue]

from here on out I will not be buying products from companies that support TPP, which pretty much subjugates entire nations to the will of corporations. Intel, Apple, Microsoft, Facebook, they'll support this awful radical anti-democratic treaty, which has so little to do with trade.

Re:Ubuntu

[nullchar]

I wouldn't call Ubuntu slim...

Re: Well

[the_humeister]

Look at the CRay results. It's a small renderer that can fit in processor cache. Look how the i5 6500, with 4 threads and lower clock, still beats an FX 8370, which has much higher clocks and 8 threads.

Re:Ubuntu

[Windowser]

Compared to Windows, it actually is

Re:Penguins are for Cows

[MobileTatsu-NJG]

... wot?

Re:It would be a news story if it DIDN'T work

[Anonymous Coward]

Skylake's HD 530 is awfully slow, it's barely above HD4600 in 4790k. Look at how bad is it in comparison to Broadwell 5775C... They had to cut it down due to power inefficiency of added SGX comparing to Broadwell. New NUC6i5SYH is slower than NUC5i5RYH due to its graphics card as well, which is a major disappointment... Skylake is a dud.

Last I heard

[rsilvergun]

Skylake was still getting bested by the gpu in the 7850k (and a 7850k combo is about $100-$150 cheaper). That said if you slap a $60 video card into an i3 rig it'll outperform the 7850k, you were only out $20 or $30 extra bucks and you had a board/cpu combo that you could put an i5/i7 in at some point, where the 7850k board was never going to take anything above an 8350. What seems to kill the 7850k's value proposition is that it needs the 2400 ram to hang with the i5 in GPU performance, and the extra cost kills it for a low end build. Putting top end ram in a low end build just seem silly...

Heat

[rsilvergun]

is the enemy of electronics. One of the things I'm wondering is if these low power (and low heat) boxes are going to last a lot longer, and eventually create a glut.

should i upgrade?

[Espectr0]

I have a third gen i5-3570k. I only use it for gaming (fps, open world third person games mostly). Should i upgrade to this 199$? Another one? Or wait? I have a nvidia 970 gtx.

Re:should i upgrade?

[mister_playboy]

You'd have to get a new motherboard and RAM to upgrade to this. Not worth it for your use.

Re:should i upgrade?

[greenfruitsalad]

according to this, the new processor would be slightly slower than your old one. plus, your old one can be easily overclocked.
http://cpubenchmark.net/compar...