Slashdot Mirror

cpubenchmark.net is these days quite useful tool for comparison. You get a rough number, which isn't applicable to all use scenarios, but is still worlds better than clock rate.

There never was a 4GHz P4 released, so it's not clear how that's even in the database. Maybe it's some engineering sample, though as someone else pointed out it's half as fast as the 3.8 Ghz P4 (the fastest released Pentium 4) so my guess it's an error. In any case, I wouldn't use if for any meaningful comparisons. Also, you might want to note that the Core i7 has a turbo boost up to 3.3 Ghz and I would have to assume that's what it's really running at during any benchmark, so to call it a 1.7 Ghz CPU is misleading at best. It's also a dual core, so to compare it to single core is also bit unfair. Perhaps the closest comparison might be the Pentium D Extreme Edition which was two hyperthreading P4's in a single chip which would be the closest thing to the i7.
http://www.cpubenchmark.net/compare.php?cmp[]=1955&cmp[]=1130

You'll see that the i7 is now only about 4-5 times faster.

Oh, yea?

An Intel 1.7Ghz i7 is TWENTY FIVE times faster than an Intel Pentium 4 4.0Ghz
http://www.cpubenchmark.net/co...

That settles it. Ignore clock speed across generations.

I don't know why this is even a story. Technology getting cheaper over time? Competition driving lower costs? Amazing!

BTW, here's a Dell for $199, so the price point isn't even new. Although, this CPU is about half as powerful (~approximate benchmark)

It's still very much the same type of device.

You're looking at the Early 2014 Haswell 11" Macbook Air. The S200E has been out for over a year now; you're not even comparing apples to apples. I'm going to roll with the mid-2013 Air even though that's still newer than the S200E and thus is still not quite fair.

How is my S200E better than a mid-2013 Macbook Air?
Price: $510 ($430 for the S200E + $80 for a 120GB SSD) vs. $899 = I paid $389 less. I can buy another S200E base unit for that price today.
Physical attributes: S200E is thicker and heavier than the Air but is still a solid aluminum chassis (other than the rubberized bottom) that doesn't tweak, crack, and break as easily as the Air; the rubberized bottom also makes the unit much less slippery overall. Servicing the S200E is also very easy. The Air might be a little lighter, but in a backpack full of other stuff, why complain about an extra half a pound, especially when it means a much sturdier unit?
CPU: The Air is a Haswell i5-4250U; the S200E is an Ivy Bridge i3-3217U; this obviously makes the Air's CPU more powerful (PassMark: 3419) than the S200E's CPU (Passmark: 2292). However, my purchase was partly based on picking a low power consumption laptop to attach to a set of solar panels, not on maximum CPU performance; that i5 is 15W TDP while the i3 is 10W TDP (1/3 lower). I don't feel bad about paying $389 less for the slower CPU though.
Storage: The S200E came with a 7mm 500GB hard drive. I upgraded it to a 120GB SATA-III SSD. My final price includes the part cost for that SSD. Its 8GB less than the Air's 128GB PCIe SSD and the performance between the two in real-world usage is probably identical (though the PCIe SSD shows better raw read speeds in simple benchmarks). Ultimately, the SSD differences are insignificant. The S200E comes with a portable slim USB 2.0 DVD-RW drive; the Air doesn't have an optical drive at all.
Ports: Both have SD card slots. Air has 2x USB 3.0 ports and I wouldn't mind having that on my S200E, but the S200E has 3 total USB ports, so I can plug in my mouse and USB 2.0 microphone and still have my USB 3.0 port free for my USB 3.0 external hard drive or flash drive when needed. Other than having more USB 3.0 ports, the Air clearly loses on ports vs. the S200E: no HDMI, no ethernet, no VGA, one less USB port. Sure, it has Thunderbolt, but Thunderbolt is useless without expensive stuff to plug into it (remember that $389 I saved? Tack on a $29 Thunderbolt network adapter add-on to that if you like.) My HDTV and 28" monitor have HDMI; does your HDTV have Thunderbolt? Nope. Does your sub-$300 28" monitor have Thunderbolt? Nope.
Input: ah, yes, Mac users love their trackpads...I hate to tell you this, but it's just a Synaptics ClickPad. The S200E has an ElanTech version of the exact same thing, and guess what? It works just as well. It detects when you're pushing to click and doesn't go haywire and move your pointer while clicking (as early PC ClickPads tended to do), it has a heap of multi-touch gestures from the factory and those gestures are all configurable, and it tracks wonderfully. The keyboard has great tactile response and even looks like it was taken directly from a Macbook Air, save the lack of cloverleaf and apple keys. The two computers are on equal footing regarding input devices. The S200E has a touchscreen as well, but fuck touchscreens.
Display: The panels in use are apparently identical; I doubt there are many choices for a thin glossy 11.6" LED LCD at 1366x768.
Conclusion: I paid $389 less. I sacrificed 1/3 of the CPU power but got 1/3 the CPU wattage back in exchange. I only got one USB 3.0 port but

You're looking at the Early 2014 Haswell 11" Macbook Air. The S200E has been out for over a year now; you're not even comparing apples to apples. I'm going to roll with the mid-2013 Air even though that's still newer than the S200E and thus is still not quite fair.

How is my S200E better than a mid-2013 Macbook Air?
Price: $510 ($430 for the S200E + $80 for a 120GB SSD) vs. $899 = I paid $389 less. I can buy another S200E base unit for that price today.
Physical attributes: S200E is thicker and heavier than the Air but is still a solid aluminum chassis (other than the rubberized bottom) that doesn't tweak, crack, and break as easily as the Air; the rubberized bottom also makes the unit much less slippery overall. Servicing the S200E is also very easy. The Air might be a little lighter, but in a backpack full of other stuff, why complain about an extra half a pound, especially when it means a much sturdier unit?
CPU: The Air is a Haswell i5-4250U; the S200E is an Ivy Bridge i3-3217U; this obviously makes the Air's CPU more powerful (PassMark: 3419) than the S200E's CPU (Passmark: 2292). However, my purchase was partly based on picking a low power consumption laptop to attach to a set of solar panels, not on maximum CPU performance; that i5 is 15W TDP while the i3 is 10W TDP (1/3 lower). I don't feel bad about paying $389 less for the slower CPU though.
Storage: The S200E came with a 7mm 500GB hard drive. I upgraded it to a 120GB SATA-III SSD. My final price includes the part cost for that SSD. Its 8GB less than the Air's 128GB PCIe SSD and the performance between the two in real-world usage is probably identical (though the PCIe SSD shows better raw read speeds in simple benchmarks). Ultimately, the SSD differences are insignificant. The S200E comes with a portable slim USB 2.0 DVD-RW drive; the Air doesn't have an optical drive at all.
Ports: Both have SD card slots. Air has 2x USB 3.0 ports and I wouldn't mind having that on my S200E, but the S200E has 3 total USB ports, so I can plug in my mouse and USB 2.0 microphone and still have my USB 3.0 port free for my USB 3.0 external hard drive or flash drive when needed. Other than having more USB 3.0 ports, the Air clearly loses on ports vs. the S200E: no HDMI, no ethernet, no VGA, one less USB port. Sure, it has Thunderbolt, but Thunderbolt is useless without expensive stuff to plug into it (remember that $389 I saved? Tack on a $29 Thunderbolt network adapter add-on to that if you like.) My HDTV and 28" monitor have HDMI; does your HDTV have Thunderbolt? Nope. Does your sub-$300 28" monitor have Thunderbolt? Nope.
Input: ah, yes, Mac users love their trackpads...I hate to tell you this, but it's just a Synaptics ClickPad. The S200E has an ElanTech version of the exact same thing, and guess what? It works just as well. It detects when you're pushing to click and doesn't go haywire and move your pointer while clicking (as early PC ClickPads tended to do), it has a heap of multi-touch gestures from the factory and those gestures are all configurable, and it tracks wonderfully. The keyboard has great tactile response and even looks like it was taken directly from a Macbook Air, save the lack of cloverleaf and apple keys. The two computers are on equal footing regarding input devices. The S200E has a touchscreen as well, but fuck touchscreens.
Display: The panels in use are apparently identical; I doubt there are many choices for a thin glossy 11.6" LED LCD at 1366x768.
Conclusion: I paid $389 less. I sacrificed 1/3 of the CPU power but got 1/3 the CPU wattage back in exchange. I only got one USB 3.0 port but

I take it back. No-one is using the top of the line atom. The other live around 1000 points passmark, so less than the core2duo. I rest my case.

Someone above compared the newest atoms to an i3. Not quite. But they are more powerful than I gave them credit. Might actually be able to out perform the old x200T workhorse.

passmark

Which uses Passmark, which is a simple corner-case number-crunching bonanza. Pure AVX2, or FMA without any real-world qualifiers, restriction or branching? Sure, we got that!

And even with that, you're still off. The performance improvement with Haswell per-core is less than 5x. See here:

http://www.cpubenchmark.net/compare.php?cmp[]=2020&cmp[]=1127

So, in the unoptimized case the performance improvement is 2-3x, and in the embarrassingly-parallel case the speedup is 4-5x. But then, if you had such an embarrassingly-parallel case, you'd just port it to OpenCL and be done with it. Haswell is for all those hard-to-optimize compute cases.

Why don't we ever read about more useful metrics, such as the amount of (floating-point) operations per second per $ of a given CPU?

Amen, brother. For example, let's take a Pentium 4, 3.0GHz and a Core i7, 3.5GHz. At the same TDP (and all cores utilized), the Core i7 is 28x more powerful than the P4. Even if we compensate the clockspeed to be equal (downclock the i7 to 3GHz), it clearly shows that something like "GHz" is today completely useless for comparing CPUs.

Why don't we ever read about more useful metrics, such as the amount of (floating-point) operations per second per $ of a given CPU?

Amen, brother. For example, let's take a Pentium 4, 3.0GHz and a Core i7, 3.5GHz. At the same TDP (and all cores utilized), the Core i7 is 28x more powerful than the P4. Even if we compensate the clockspeed to be equal (downclock the i7 to 3GHz), it clearly shows that something like "GHz" is today completely useless for comparing CPUs.

Not really, since it still gets beaten out by last generation's quad core i7 3770 @ 3.4GHz.

http://www.cpubenchmark.net/cp...

Why don't we ever read about more useful metrics, such as the amount of (floating-point) operations per second per $ of a given CPU?

Because the target market for this thing doesn't consider that a useful metric, and never has.

For some years now (at least back to the P4 era, if memory serves), Intel has always offered the mad-crazy-overclocker-must-go-faster-edition CPU at the top of their (desktop, sorry Xeon buyers!) price list, usually ~$1,000. This part is always an astonishingly poor value, unless what you want is the fastest x86 money can buy. Most of them go to gamer e-peen setups, they may sell some to compute customers who have some pathologically hard-to-parallelize problem and thus need the fastest single threaded performance they can get, rather than more cores with lower performance per thread but far lower cost.

If you are actually shopping for CPUs, you probably want something like CPUboss, or CPUbenchmark which allows you to do fairly easy comparisons of performance/price (albeit for performance as measured by one or more general benchmarks, if your workload is somewhat atypical, your mileage may vary).

A comparison amongst users in typical rigs, rather than in review-site benchmark rigs.
i7-860 1217 / 5110 (812 samples, single threaded / overall)
FX-8350 1512 / 9049 (3149 samples, single threaded / overall)

The AMD in question is winning against the Intel in question in single threaded, and winning greatly in multi-threaded. However this AMD chip, at $200, is not really what the GP was talking about. He was talking about ~$150 APU's that also saves him money on a video card.

Comparison of the i7-860 vs the A10-6800K

i7-860 1217 / 5110 (812 samples, single threaded / overall)
A10-6800K 1555 / 5006 (205 samples, single threaded / overall)

A comparison amongst users in typical rigs, rather than in review-site benchmark rigs.
i7-860 1217 / 5110 (812 samples, single threaded / overall)
FX-8350 1512 / 9049 (3149 samples, single threaded / overall)

The AMD in question is winning against the Intel in question in single threaded, and winning greatly in multi-threaded. However this AMD chip, at $200, is not really what the GP was talking about. He was talking about ~$150 APU's that also saves him money on a video card.

Comparison of the i7-860 vs the A10-6800K

i7-860 1217 / 5110 (812 samples, single threaded / overall)
A10-6800K 1555 / 5006 (205 samples, single threaded / overall)

Or this...
Chart to sort by Passmark rank
and check the CPU one as well.

Ha ha, your setup sounds remarkably similar to mine :)

After evaluating the FitPC and a bunch of mini-ITX stuff for work, I ended up getting a (secondhand) nVidia ION / Atom shoebox PC to migrate my 24x7 home server to. With the nVidia drivers, it still makes a decent and responsive desktop with compiz-fusion. The only time I notice that it's actually running on a poky nettop is on the busier flash-crappy web pages.

Ended up getting an AMD-based Toshiba laptop for my wife to replace the crappy iBook I bought her to break her of her Mac habit. It was a big, cheap 17" desktop-replacement deal with ATi HD3500 something graphics. Unfortunately, she got hooked on Windows, but Toshiba abandoned driver support for it, so when we migrated off Vista, the built-in MS drivers do decent 3D acceleration but no 2D acceleration. So it stutters horribly doing any kind of fullscreen video like youtube or Netflix, which is mostly what it does now that she's finished her dissertation.

Haven't really tried to pay attention much at PCs these days, but I find http://www.videocardbenchmark.net/ and http://www.cpubenchmark.net/ indispensible nowadays when I come across a new system at work and I want to know how much it sucks based on the alphabet soup in its name, so I can wield it with the appropriate amount of swagger compared to what my peers are packing.

Of course, you could then look at this benchmark is more indicative of what people actually run: http://www.cpubenchmark.net/singleThread.html

$80...$100 The highest benchmark score for Intel is 1,988 with the G3430, and the highest benchmark score for AMD is 1,385 with the AMD A8-5600K
$100..$120 The highest benchmark score for Intel is 1,797 with the i3-3240 (G3430 better buy), and the highest benchmark score for AMD is 1,526 with the FX-4350.
$120..$140 The highest benchmark score for Intel is 1,859 with the i3-3245, and the highest benchmark score for AMD is so far down the list I got bored.
same trend continues in every bracket

Sure, if you ignore either the price or the performance you can imagine your statement to be true.

I'm curious as to exactly what 2 watt processors can compare to a 3.2 GHz P4 from 2003

It's not easy to compare, and I had to jump around benchmarks a bit, but some of the recent SoCs look to be ballpark with the P4.

x86 Atom Z2760 Vs ARM
http://www.notebookcheck.net/SoC-Shootout-x86-vs-ARM.99496.0.html

Atom D510 Vs P4
http://www.tomshardware.com/reviews/atom-d510-pentium-4-nettop,2649-10.html

Atom D510 Vs Atom Z2760
http://www.cpubenchmark.net/midlow_range_cpus.html

Today they are still the value king.

The best value i-series Intel shows up at 37.68.
The best value A-series AMD shows up at 44.95
The best value FX-series AMD shows up at 55.66.

That best value i-series is both more expensive and slower than the best value A-series, so no mixing apples and oranges here. Intel is still getting slaughtered on value, and thats before factoring in their more expensive motherboards.

BULLSHIT.

An Athlon 2400+ has a CPUMark of 357.

Athlon or Athlon XP? (your mark seems to be from the latter)

An A7 (iPhone 5S) has a CPUMark of 35274.

You workstation is slow by any standard. Modern SoCs are much faster than you imagine. Even your Geforce 4 is so abysmal slow compared to modern mobile GPU that you assertion is simply laughable.

Welcome to reality. Phones annihilate you ancient workstation and blow it in the middle of next week.

Indeed, this page gives the same results too. It dwarfs even the current state-of-the-art workstation class CPU, Xeon E5-2687W, which could only net 14630 CPUMarks!

BULLSHIT.

An Athlon 2400+ has a CPUMark of 357.

Athlon or Athlon XP? (your mark seems to be from the latter)

An A7 (iPhone 5S) has a CPUMark of 35274.

You workstation is slow by any standard. Modern SoCs are much faster than you imagine. Even your Geforce 4 is so abysmal slow compared to modern mobile GPU that you assertion is simply laughable.

Welcome to reality. Phones annihilate you ancient workstation and blow it in the middle of next week.

Indeed, this page gives the same results too. It dwarfs even the current state-of-the-art workstation class CPU, Xeon E5-2687W, which could only net 14630 CPUMarks!

BULLSHIT.

An Athlon 2400+ has a CPUMark of 357.

Athlon or Athlon XP? (your mark seems to be from the latter)

An A7 (iPhone 5S) has a CPUMark of 35274.

You workstation is slow by any standard. Modern SoCs are much faster than you imagine. Even your Geforce 4 is so abysmal slow compared to modern mobile GPU that you assertion is simply laughable.

Welcome to reality. Phones annihilate you ancient workstation and blow it in the middle of next week.

Indeed, this page gives the same results too. It dwarfs even the current state-of-the-art workstation class CPU, Xeon E5-2687W, which could only net 14630 CPUMarks!

This is why a mid-range Intel part (Say, an Intel Core i5-4670K) can handily (and significantly) beat AMD's top-of-the-line desktop CPU (An FX-8350)

Really? CPU Benchmarks says

  i5-4670K - 7531

  AMD FX-8350 - 9091

A comparable Intel chip would have to be closer to i7-3820, not your i5. Perhaps your benchmarks are a little crappy?

http://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-8350+Eight-Core&id=1780

http://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-4670K+%40+3.40GHz&id=1921

Anyway, AMD is far more $$$ efficient for typical desktop. Yes, including any thermal envelope differences.

Finally some facts.

This is why a mid-range Intel part (Say, an Intel Core i5-4670K) can handily (and significantly) beat AMD's top-of-the-line desktop CPU (An FX-8350)

Really? CPU Benchmarks says

  i5-4670K - 7531

  AMD FX-8350 - 9091

A comparable Intel chip would have to be closer to i7-3820, not your i5. Perhaps your benchmarks are a little crappy?

http://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-8350+Eight-Core&id=1780

http://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-4670K+%40+3.40GHz&id=1921

Anyway, AMD is far more $$$ efficient for typical desktop. Yes, including any thermal envelope differences.

Finally some facts.

I initially wrote this reply to another post, but only noticed that he was trolling right before submitting. I think my experiences through the years could be useful for a budget-conscious gamer building his own rig, so instead of deleting it I leave it here. Here are a few points to consider:

• * Buy quality wear parts, mid-range performance parts, upgrade only what you need for a new iteration. I have done this for 15 years, and have always been able to play the latest games at decent settings. My upgrade costs have been pretty stable at around $800 on average every two to three years (running on three years with my current rig, upgrading these days), giving me solid performance for less than $300/year. The quality brand 600W PSU I got five years ago, for instance, will still be present in my new rig, and maybe even the next one after that as long as it's cable compatible with new components.
• * Don't pay the huge premium on the latest GPU/CPU, you can get mid-range alternatives with 80% of the performance at 1/2 the price. Check out CPU and GPU charts (sort by rank), or actual game performance reviews to find good alternatives. No current game *needs* an i7 or a GTX 7XX video card to get great fps at high settings on a single 1920x1200 monitor. The ROI drops off in a ridiculous rate if you go for the newest hardware. It won't even pay off significantly in performance, see the next point.
• * Realise that all games are made to run very well on computers far below high-end. At the end of my 2.5-ish year upgrade cycle I might lose out on some minor eye-candy in the latest games. I also get to laugh at the tools who post in forums with comparison screenshots discussing whether the shiny they were able to turn on and run at 90 FPS with their two new $1019 video cards is even *visible in comparison screenshots*. Seriously.
• * Don't be stupid about what you really need. A 512GB SSD, for instance, is completely ridiculous in a budget gaming rig. A 120GB or even a 90GB one (as I have in my current rig and will keep for the next iteration) will hold a couple of OS'es and the 3-4 games you currently play. Just use mklink (or Steam Mover, which automates this for any game or application, not just Steam ones) to swap them in from your humongous bulk storage spinning drive. It will not degrade your gaming experience at all. If you pay a huge premium to avoid a three-minute coffee break every two weeks to shift in a new game that's your choice, but don't complain about it :)

• * Spend a few dollars more to get a mainboard with an automatic overclocking analyzer along with a a $30 aftermarket silent CPU cooler. I've never bothered to overclock manually, but if the computer does the work for me, I'll take the 10-30% stable CPU performance increase you can get "for free". A $40 premium for this feature and a cooler is *a lot* cheaper than actually buying a chip specced from the factory for those increased speeds. Even my current el-cheapo ASRock board did this three years ago.

• * Open up your case and clean your computer thoroughly once in a while. Remove all dust from air channels in cooling ribs and the video card. This one might be obvious, but very few people actually do it. The computer will run cooler, be more quiet and last longer.

I completely agree with you about brand name components. After two bad experiences with a relatively cheap video card and a mainboard from unknown manufacturers I only buy brands myself (including EVGA for video cards and ASRock for mainboards).

Yes, it does take a bit of research. No, you don't really have to spend more than a couple of hours on it if you don't want to.

This is why a mid-range Intel part (Say, an Intel Core i5-4670K) can handily (and significantly) beat AMD's top-of-the-line desktop CPU (An FX-8350)

Really? CPU Benchmarks says

  i5-4670K - 7531
  AMD FX-8350 - 9091

A comparable Intel chip would have to be closer to i7-3820, not your i5. Perhaps your benchmarks are a little crappy?

http://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-8350+Eight-Core&id=1780

http://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-4670K+%40+3.40GHz&id=1921

Anyway, AMD is far more $$$ efficient for typical desktop. Yes, including any thermal envelope differences.

This is why a mid-range Intel part (Say, an Intel Core i5-4670K) can handily (and significantly) beat AMD's top-of-the-line desktop CPU (An FX-8350)

Really? CPU Benchmarks says

  i5-4670K - 7531
  AMD FX-8350 - 9091

A comparable Intel chip would have to be closer to i7-3820, not your i5. Perhaps your benchmarks are a little crappy?

http://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-8350+Eight-Core&id=1780

http://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i5-4670K+%40+3.40GHz&id=1921

Anyway, AMD is far more $$$ efficient for typical desktop. Yes, including any thermal envelope differences.

Yeah...well no, you might want to look up the price/core cost vs AMD and Intel, then you'll quickly see AMD tromps all over it. And really with the Vishera cores, you're seeing a negligible loss in real world performance. The only place where Intel beats AMD in cost-per-core is with the celery(celeron) line.

You do realize that your chart is heavily skewed toward CPUs which aren't current production, right? I mean, the top 11 SKUs on that chart range in price from $11 to $22 and can only be purchased from random 3rd parties on Amazon at fire sale prices. You have to get to the 12th SKU in order to find something available at retail (from New Egg)... and it's an Intel product. #13 is also an out of production unit, while #14 is another Intel product. #15 is an honest to god current production AMD chip, though.

While I won't accuse you of lying with data (I've done quick research that appeared to support my conclusion and ended up with egg on my face, as well) the point you're trying to make (other than that Celeron beats AMD in price/performance) is totally undermined by your supporting data.

It is also significantly slower buck for buck in real life workloads.

Yeah...well no, you might want to look up the price/core cost vs AMD and Intel, then you'll quickly see AMD tromps all over it. And really with the Vishera cores, you're seeing a negligible loss in real world performance. The only place where Intel beats AMD in cost-per-core is with the celery(celeron) line.

Again sure if you took say the Pentium D EE and used a bad enough cooler you could get up to 6GHz plus, but how much useful work would you get for that heat? To use a /. car analogy it doesn't matter if you have an engine that can take a high redline if the transmission doesn't turn those revs into speed. I mean look at this comparison using the Pentium 4 661, which was one of the fastest P4 chips they made that wasn't an extreme. You see that C60 that scores 2 marks higher? That is a 1GHz netbook chip that uses a grand total of...9 watts. A 9 watt netbook chip scores higher than one of the fastest MHz P4s that was ever made.

So I'm sorry but reaching 10GHz really wouldn't help when the IPC is THAT bad, if the passmark stayed relative to speed you'd be scoring around 1300 marks at 10Ghz, which would be equal to a 4 year old Athlon X2 at a blistering...2.5Ghz. This is the same problem that AMD is now suffering from, their BD/PD design just doesn't have good IPC and cranking up the speed is just gonna bring it from bad to okay, and the amount of heat you'll have to deal with really isn't worth it.

Again sure if you took say the Pentium D EE and used a bad enough cooler you could get up to 6GHz plus, but how much useful work would you get for that heat? To use a /. car analogy it doesn't matter if you have an engine that can take a high redline if the transmission doesn't turn those revs into speed. I mean look at this comparison using the Pentium 4 661, which was one of the fastest P4 chips they made that wasn't an extreme. You see that C60 that scores 2 marks higher? That is a 1GHz netbook chip that uses a grand total of...9 watts. A 9 watt netbook chip scores higher than one of the fastest MHz P4s that was ever made.

So I'm sorry but reaching 10GHz really wouldn't help when the IPC is THAT bad, if the passmark stayed relative to speed you'd be scoring around 1300 marks at 10Ghz, which would be equal to a 4 year old Athlon X2 at a blistering...2.5Ghz. This is the same problem that AMD is now suffering from, their BD/PD design just doesn't have good IPC and cranking up the speed is just gonna bring it from bad to okay, and the amount of heat you'll have to deal with really isn't worth it.

You forgot the cost of replacing RAM.

Also: Intel Celeron G1610 Ivy Bridge 2.6GHz, Dual core w/heatsink & fan - ~$50 Intel Desktop Board Classic Series MicroATX DDR3 1333 LGA 1155 Motherboard - ~$50 Total - ~$100 (free shipping)

That Celeron scores 2621 on the same benchmark site that you linked.

I think you're terribly wrong. A T3500 uses Socket 1366 Xeons from 2009. The most powerful one offered with that workstation is Nehalem EP (1st gen i7) although I very much doubt they sell that out of the box to financial companies anyway, which probably means he got the standard one that comes with a Bloomfield (Core2Quad).

Let's be generous and assume he got a fully maxed out T3500 and it came with a i7 though, just for kicks.

http://www.cpubenchmark.net/cpu_lookup.php?cpu=Intel+Xeon+X5570+%40+2.93GHz&id=1302

The AMD fx-8350 kicks it's backside and is $190, and most importantly, you won't need to throw everything away when it's time to upgrade. Spending an extra $100 now and saving $600 on buying new stuff over the course of the next 3 upgrades is how I like to work. That's why my current system (3930k, 32GB RAM 12TB of HDDs) was affordable on my student budget, because I already owned everything except the CPU and mob when I went ot buy them, and I just get bits and bobs one by one.

Thanks, and for those that say "Oh and E-Series doesn't stack up against a COre2Duo"? Here is the Passmark scores which the E350 scores 776, the Intel C2D T2400 at 1.83GHz scores a 789, so pretty damned close to the first gens. If you can score an E450 they score neck and neck with the 2GHz T2450 so again its neck and neck only the price is a HELL of a lot different.

And to the AC that says "I like AMD"? I like bang for the buck and hate corporate douchebaggery, intel STILL rigs their compiler (while AMD uses a fork of GCC and hands out the code) and if you'll show me where I can get a C2D WITH the heatsink AND the motherboard for $70 shipped? be happy to consider it. With the E350 you can take an old Pentium 4 box and turn it into a power sipper with better performance for less than $110, and that is with you buying a PCI to IDE adapter if the system still has IDE drives, if it has SATA you can do it for less than $95. I'm sorry but you can't beat that, you save money, save cooling, and save power, while getting better performance...what's not to like?

Looks like you are in the ballpark:

Intel Core2 Duo T5500 @ 1.66GHz - score 882
AMD E-350 APU - score 881

I run mine completely fanless and passively cooled by its (rather massive) heatsink. It gets hot under load, though not enough to trip the thermal protection scaler. Furthermore, it's silent except for capacitor whine.

As I noted above, the platform is marginal with regard to crypto filesystem throughput even on the lowest possible settings that various crypto frameworks offer. One crucial factor that gives it an edge over an Atom-platform unit is that the system supports up to 8 GB of RAM vs. Atom's 4 GB.

I would be completely satisfied with my E-350 if it supported AES-NI, which would bump the crypto throughput into the 100+ megabyte per second range.

Put another way, if you go back 10 years from the P4 3.8GHz in 2004, you are at the Pentium 75 in 1994. I don't even know where to find a single benchmark to compare the two! The performance difference would be more like 20x, not 2.5x. Intel went from the P75 to the P200 in 20 months, and that was a bigger increase in single-core performance than we've seen in the last decade.

Increases in IPC are very small compared to what we enjoyed as clock speed ramped up.

The dual core Zacate E-450 at 1,65 GHz reaches a score of 806 in the Passmark CPU test.
Here is a comparison to the bulldozer: http://www.cpubenchmark.net/cpu.php?cpu=AMD+E-450+APU&id=250
If we are optimistic on behalf of the PS4 and assume that the 8 cores are four times as fast and there is also a boost from higher clock speed, we might get something like a Passmark score of 4000. Which brings us to the level of a AMD FX-4100, $ 105 on Newegg right now.
A matching mainboard: at least $50.
If you look at the GPU, it is more like a HD7850, around $200.
RAM is a bit hard to compare, as PCs don't have GDDR5 main memory. 8 GByte of "slow, ordinary" DDR3 are around $60.
HDD: Sony did not specify the sitze, but lets take a 3.5", 1TByte for comparison. Around $80.
A halfway decent case: $60.
and finally a power supply, 400W or more: $50.

So far we are at about $600. Add a keyboard & mouse, and we approach $650. Of course that PC is way more versatile than the PS4, so I would prefer it over the console (I might also get a different CPU, the FX-4100 was mainly for comparison).

I still feel a bit of sore, so I'll chip in the discussion.
Few months ago I read several good reviews about the Zboxes as HTPC, so I bought one, together with 4Gb of RAM and a 160Gb SSD. The GPU was a ION and no, there were no chances to have a decent XBMC experience nor playing 720i videos (1080p? don't even think about it) without having very unpleasant "hiccups" here and there in the playback. Forget about any online streaming with more than 360p resolution (average YouTube videos were enough to put the thing on its knees). No, desktop effects were not on the way (i.e. barebone XFCE). It shouldn't have been a surprise, since it was sporting a crappy Atom D525 processor 1.8GHz, and I blame myself for having bought the positive reviews despite the terrible hardware specs.
After two frustrating weeks of tweaking trying to squeeze more juice out of it (Xorg.conf, VDPAU, Nvidia drivers...), I've sent it back and with pretty much the same amount of money, I've tried one of the cheap solutions I've found on the XBCM forum.
I got a i3@3.1GHz machine that's able to run smoothly Black Mesa, SteelStorm and TF2.
The box slashvertized here has a Celeron 847 instead of the Atom, but CPU performances are equally poor, so don't even think about games (or at least nothing more complex than Gchess).

The rather trivial moral of the story is that if you want a powerful machine you should buy one, and not waste your time with toy computers.
Be wise with your money and never look back.

Looking at the performance charts at Passmark the top still looks dominated by AMD. Intel has ramped up performance, but it's ramped up price even more.

Intel makes chips with more than 8 cores.

10 core Xeon: http://ark.intel.com/products/53580/Intel-Xeon-Processor-E7-8870-30M-Cache-2_40-GHz-6_40-GTs-Intel-QPI

Granted, it's incredibly expensive (as you point out) and I've only seen them in blade applications. But, they do make them. It's also worth pointing out that on the whole, one intel core gives far superior performance than one AMD core of the same clock speed (see http://www.cpubenchmark.net/high_end_cpus.html). Moreover, Intel's hyperthreading can be of a huge help, if your application profile fits.

Measuring $/core or $/CPU Cycle is not a very accurate way to gauge price/performance.

HUGE!!!

why don't you look at something like this?

Not sure if you are joking.

The Turion 1.5 GHz is more powerful than an Atom D525 1.8 GHz.

http://www.cpubenchmark.net/cpu.php?cpu=AMD+Turion+II+Neo+N40L+Dual-Core
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Atom+D525+%40+1.80GHz

The Atom also won't support ECC or virtualization instructions. That tiny box only allows a single drive, so forget about RAID. If the tiny box dies, you will need a screwdriver to move the hard drive to a spare tiny box and it will be a slow process; if the hard drive dies, the server is down (no RAID).

The HP MicroServer has a card slot that can take a system management card. Also, if a data center is already buying HP kit, maybe they will want to buy an HP MicroServer instead of an "E-Box" from a Yahoo store.

But other than that, yeah I guess one is as good as the other.

HUGE!!!

why don't you look at something like this?

Not sure if you are joking.

The Turion 1.5 GHz is more powerful than an Atom D525 1.8 GHz.

http://www.cpubenchmark.net/cpu.php?cpu=AMD+Turion+II+Neo+N40L+Dual-Core
http://www.cpubenchmark.net/cpu.php?cpu=Intel+Atom+D525+%40+1.80GHz

The Atom also won't support ECC or virtualization instructions. That tiny box only allows a single drive, so forget about RAID. If the tiny box dies, you will need a screwdriver to move the hard drive to a spare tiny box and it will be a slow process; if the hard drive dies, the server is down (no RAID).

The HP MicroServer has a card slot that can take a system management card. Also, if a data center is already buying HP kit, maybe they will want to buy an HP MicroServer instead of an "E-Box" from a Yahoo store.

But other than that, yeah I guess one is as good as the other.

Anyway, care to provide a citation about the "terrible floating point performance"? As far as I know AMD Opterons has much better multicore floating point performance / dollar than Intels.

He's probably referring to the Buldozers which do have terrible FP performance, primarily because two "cores" share a single FPU.

For example, while the FX-8120 gets about 77% of the score the i7 2600k gets in PassMark, in a very floating-point heavy, multi-threaded rendering application the 8120 gets about 51% of the 2600k.

I do not think I am debunking anyone; I am conversing. When I debunk, it's much bloodier.

This sums up the other link:

http://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-8150+Eight-Core

Let's say 8150 is slower , 20-30% on single threading (I am not saying it's true, I am saying people say it) than intel's chip. Is single threaded the normal use-case or is that a competitive gamer thing (obsession) where the only thing separating you from your opponent is not skill or strategy but CPU speed on a single thread.?

Mostly, in my life, I am using more than one thread. I am doing a number of things at once. The OS wants one (or more). My programs all want as many as I've got. Even people who aren't working with IDEs and rendering applications are still, say, listening to music and watching a video and all this kind of thing all at once.

Intel's chip costs more, are slower except on single threaded applications, Intel is evil. I can OC the chip easily and have a nice stable system that is just as fast for zero extra dollars on a single threaded application.

But the overall thing to not lose sight of is -the chips are stupifyingly fast . We can look at CPU bench marks all day but mostly they sit idle waiting for our I/O to hurry up.

Not arguing here. Just observing and thinking aloud.

From passmark's score for high end CPUs:

http://www.cpubenchmark.net/high_end_cpus.html

AMD FX-8140 Eight-Core

score: 7,133 $169.00

then about 20 CPUs down the ranked list :

Intel Core i5-2500K @ 3.30GHz

score: 6,580 $217.99*

http://www.cpubenchmark.net/high_end_cpus.html

AMD Phenom II X6 1045T - scrore 4999
Intel Xeon X3450 @ 2.67GHz - score 5223

X6 costs with MB $90.

I buy and use AMD simply because it is the best value for the money

http://www.cpubenchmark.net/cpu_value_available.html

Microcenter or Fry's throws is also cheap motherboard. For $90 you can get top performing AMD CPU and motherboard. My 4 core AMD supports 3 way multiseat and runnning 4 X11 sessions on Ubuntu 12.04 just fine. Match that Intel.

The A series have terribly slow cores that aren't much better than Atom cores.

Citation needed. The cores are based on the Mobile Phenom II architecture Llano. In Benchmarks the A6-3650 ranks around a Phenom II X4. While this is not top of the line, it's pretty decent power.

Atoms aren't even in the same ballpark. Perhaps you are thinking of the E-350, which is in the Atoms league but with better graphics? Try even finding Atom motherboard supporting SATA3, USB3 and 16GB RAM (many of the FM1 motherboards support up to 64GB RAM!)

FM1 is a dead end

So what? Intel switches sockets more often than its underwear. In this price category, upgradeability doesn't matter much. These kind of machines get built, used, and replaced. Never upgraded.

Even the i3 stomps the A8's wimpy cores.

... but you'd have to buy a graphics card. Intel isn't exactly known to have snappy graphics. Depending on your workload, more cores might be better. I was surprised, though, to see that motherboards for Intel CPU have become much cheaper. Last time, I checked most motherboards for Intel CPUs were massively more expensive. Coupled with a more expensive CPU, this made the platform uninteresting for budget computers.

A year ago, the competing i3s were slower too and the price of the motherboards were higher. Back then the An series were much better bang/buck.

The modern i3s are much faster (especially per core). Still, the cheapest i3 I can find (i3-2120) is 112.90€. That's 26€ more expensive than the A6-3650 I quoted. To have an equivalent motherboard, I'll stay within brand (Gigabyte) and set as conditions 4 RAM slots (as 4x4GB is cheapest today), SATA3, USB3. The cheapest here is GIGABYTE GA-B75M-D3H at 83.90€, which admittedly is 16€ cheaper and 2x8GB kit RAM DDR3-1333 from ADATA (67.98€) which is identical to my original quote. Overall cost of the i3 platform is thus indeed just 10€ more expensive. As said, this is mainly because motherboard prices became much lower than in the past. So, yes, you pay 4% more, for a 17% CPU power increase (based on passmark value). In the graphics department, the view is much less rosy (again, according to passmark values): the HD 6530D in the AMD A6-3650 delivers more than double the performance of the Intel i3 HD 2120. You'll have to compensate that with a graphics card, which costs money and east power and most likely has an additional fan and is thus louder.

One needs to look at the complete platform and decide what is more important. I guess the budget conscious gamer would be better off with the i3/Graphics card combo. The budget conscious occasional gamer, is probably better of with the AMD An processors. The budget conscious office dweller: i3 with integrated graphics.

a lot of buyers were very disappointed in the performance once purchased.

I wasn't and most reviews I read were actually quite positive. You don't expect i5 or i7 performance from these kind of CPUs. What I got was silence, enough power to have headroom and decent graphics all for a decent price.