The Haswell Refresh Processors

As a species that thrives on getting things done quicker or cheaper, a processor refresh should be welcomed. For the refresh, Intel is launching a total of 42 new SKUs, 24 of which are desktop but not all of them will make it into the land of retail. Prices for batches of a thousand will range from $42 to $303 on the desktop and $75 to $434 in mobile. Cutting to the chase, the new desktop SKUs are:

Intel Haswell Refresh CPU List (Desktop)
New SKU Cores  Base Turbo L3 Cache IGP TDP / W Intel
Celeron G1840 2 / 2 2800 - 2 MB HD (Haswell) 53 $42
Celeron G1840T 2 / 2 2500 - 2 MB HD (Haswell) 35 $42
Celeron G1850 2 / 2 2900 - 2 MB HD (Haswell) 53 $52
Pentium G3240 2 / 2 3100 - 3 MB HD (Haswell) 53 $64
Pentium G3240T 2 / 2 2700 - 3 MB HD (Haswell) 35 $64
Pentium G3440 2 / 2 3300 - 3 MB HD (Haswell) 53 $75
Pentium G3440T 2 / 2 2800 - 3 MB HD (Haswell) 35 $75
Pentium G3450 2 / 2 3400 - 3 MB HD (Haswell) 53 $86
Core i3-4150 2 / 4 3500 - 3 MB HD 4400 54 $117
Core i3-4150T 2 / 4 3000 - 3 MB HD 4400 35 $117
Core i3-4350 2 / 4 3600 - 4 MB HD 4600 54 $138
Core i3-4350T 2 / 4 3100 - 4 MB HD 4600 35 $138
Core i3-4360 2 / 4 3700 - 4 MB HD 4600 54 $149
Core i5-4460 4 / 4 3200 3400 6 MB HD 4600 84 $182
Core i5-4460S 4 / 4 2900 - 6 MB HD 4600 65 $182
Core i5-4590 4 / 4 3300 3700 6 MB HD 4600 84 $192
Core i5-4590S 4 / 4 3000 3700 6 MB HD 4600 65 $192
Core i5-4590T 4 / 4 2000 - 6 MB HD 4600 35 $192
Core i5-4690 4 / 4 3500 3900 6 MB HD 4600 84 $213
Core i5-4690S 4 / 4 3200 3900 6 MB HD 4600 65 $213
Core i5-4690T 4 / 4 2500 - 6 MB HD 4600 45 $213
Core i7-4785T 4 / 8 2200 - 8 MB HD 4600 35 $303
Core i7-4790 4 / 8 3600 4000 8 MB HD 4600 84 $303
Core i7-4790S 4 / 8 3200 4000 8 MB HD 4600 65 $303
Core i7-4790T 4 / 8 2700 - 8 MB HD 4600 45 $303

Every new desktop processor is a 100-200 MHz upgrade from the one named xx2x lower but at the same price point. Back when the initial processor price lists were announced, I made different comparisons from Celeron to i7 depending on what was relevant. For example, in the Celeron category:

Haswell Celeron
  Celeron G1840
2.8 GHz 
Celeron G1830
2.8 GHz 
CPU Speed 2.8 GHz 2.8 GHz
Cores 2 2
Threads 2 2
L2 Cache 512 KB 512 KB
L3 Cache 2048 KB 2048 KB
IGP HD (Haswell) HD (Haswell)
IGP Frequency ? 1050
TDP 53W 53W
Price at Launch $42 $52

Here we already have a Celeron processor with the exact same speed as one of the Haswell Refresh SKUs, but because it has been on the market for longer it is actually cheaper while providing the same specifications.

At the top end i7, the i7-4790 should actually be a replacement for the i7-4771 which was released in September 2013:

Haswell Core i7
  Core i7-4790
Core i7-4771
CPU Speed 3600 MHz / 4000 MHz 3500 MHz / 3900 MHz
Cores 4 4
Threads 8 8
L2 Cache 1024 KB 1024 KB
L3 Cache 8192 KB 8192 KB
IGP HD 4600 HD 4600
IGP Frequency 350 / 1200 MHz 350 MHz / 1200 MHz
TDP 84W 84W
Price at Launch $303 $314 (OEM) / $320 (box)

By releasing new mainstream processors into the desktop market, as well as the 9-series chipsets, Intel has afforded the system integrators another chance to sell newer systems and not to stagnate with older models.

For most of the performance enthusiasts reading this review, the question you might well be asking yourself relates to the Haswell Refresh overclockable CPU. At this point in time it is clear that the i5-4670K and i7-4770K models do not have refresh counterparts, both of which are also more marketable to system integrators as well as desired by home users. There is a simple explanation to this.

Nothing fundamentally changes with these refresh processors. It is the same design underneath. Now that Intel has been making Haswell wafers, the process is slightly more efficient over time (ironing out production) and the processor bins can all be shifted. If Intel were to release new overclockable CPUs that were a simple speed bump but ultimately had no performance gain of the current overclocking CPUs, there might be a backlash. 

To that end, Intel is going to release ‘Devil’s Canyon’ in due course. Devil’s Canyon has no official SKU name yet (i7-4970K or i7-4770X are my best guesses) but it was announced back in March as having an improved package and better thermal interface material. It will probably be the same silicon underneath as the i7-4770K due to the nature of the refresh, perhaps with a small design tweak that does not require new masks. We will review the new overclocking processor when we receive a sample from Intel, hopefully under something suitably extreme.

Today’s Review: The Intel Core i7-4790, i5-4690 and i3-4360

Back when Haswell was launched, we had only the chance to test the top line overclocking processor, the i7-4770K. This time around I was thankful to get the opportunity to expand our testing to the i5 and i3 segments of the processor line to find where they each stand in terms of performance. For comparison points we have the i7-4770K from our launch review, an i7-4765T from our ASRock M8 review, and an i3-4330 we picked up along the way.

The Intel Haswell Refresh Review
  Core i7-4790
Core i7-4770K Core i7-4765T Core i5-4690
Core i3-4360
Core i3-4330
Price $303 $340 $303 $225 $149 $140
Launch 5/2014 6/2013 6/2013 5/2014 5/2014 9/2013
Cores / Threads 4/8 4/8 4/8 4/4 2/4 2/4
3600 3500 2000 3500 3700 3500
4000 3900 3000 3900 - -
IGP HD 4600 HD 4600 HD 4600 HD 4600 HD 4600 HD 4600
350 350 350 350 600 350
IGP Turbo 1200 1250 1200 1200 1150 1150
L3 Cache 8 MB 8 MB 8 MB 6 MB 4 MB 4 MB
TDP 84 W 84 W 35 W 84 W 54 W 54 W

Assuming all goes as planned, the small increases in frequency should provide a ~3% increase in benchmarks compared to the older Haswell processors.

Multi-Core Enhancement Isn’t Much Of An Issue

CPU benchmarking can be a bit of a nightmare in the current climate, especially on Intel platforms where turbo modes are possible. Back in 2012 I wrote an extensive news piece about Multi-Core Enhancement titled ‘The Debate About Free MHz’, wherein it is not actually Intel that set the CPU frequency for the processors but the motherboard. Where Intel might have a 3500 MHz base frequency and a 3900 MHz turbo frequency, a motherboard can detect that setting and override it for 3900 MHz on all cores by default. One motherboard went as far as increasing the multiplier +1 over the Intel specifications. The upshot of this is in multi-threaded performance benchmarks, where due to the extra frequency the final result is higher than Intel specifications. MultiCore Enhancement also means that the same CPU in different motherboards will give different results out of the box.

Most motherboard manufacturers that use this tactic (MSI, ASRock, GIGABYTE and ASUS all have this feature), as far as I can tell, limit the option to overclockable SKUs, and it is only implemented when XMP is implemented.

On top of this, Windows can also be a culprit. In different power modes, users can specify the percentage of maximum CPU frequency depending on what power mode is in play. This relates more to portable devices, but it is configurable with desktop systems as well.

For our review, as we only have one overclockable CPU in the test, we can adjust appropriately. For our testing we actually run SYSMark 2014 as our first benchmark, as the install prefers a fresh OS image to work with. This generates a consistent power profile which we maintain for the rest of the benchmarks. Had we run SYSMark mid-way through our testing period, it may have ended up with variable results, so we aim to keep our environment consistent and maintainable.

Test Setup

We would like to thank the following companies for providing test-bed support:

Thank you to OCZ for providing us with PSUs and SSDs.
Thank you to G.Skill and ADATA for providing us with memory kits.
Thank you to Corsair for providing us with an PSU,  CLC and memory.
Thank you to ASUS for providing us with the AMD HD7970 GPUs and some IO Testing kit.
Thank you to MSI for providing us with the NVIDIA GTX 770 Lightning GPUs.
Thank you to Rosewill for providing us with keyboards.
Thank you to ASRock for providing us with the 802.11ac wireless router for testing and other components.

Test Setup
Processor Intel Core i7-4790 ES (4C/8T, 3.6 GHz to 4.0 GHz)
Intel Core i5-4690 ES (4C/4T, 3.5 GHz to 3.9 GHz)
Intel Core i3-4360 ES (2C/4T, 3.7 GHz)
Motherboards ASRock Z97 Extreme6
ASUS Z97-Deluxe
Cooling Corsair H80i
Thermalright TRUE Copper
Power Supply OCZ 1250W Gold ZX Series
Corsair AX1200i Platinum PSU
Memory G.Skill RipjawsZ 4x4 GB DDR3-1866 8-9-9 Kit
ADATA XPG 2x8 GB DDR3L-1600 9-11-9 Kit
Memory Settings 1600 9-11-9-27 1T tRFC 240
Video Cards MSI GTX 770 Lightning 2GB (1150/1202 Boost)
ASUS HD7970 3GB (Reference)
Video Drivers Catalyst 13.12 WHQL
NVIDIA Drivers 335.23 WHQL
Hard Drive OCZ Vertex 3 256GB
Optical Drive LG GH22NS50
Case Open Test Bed
Operating System Windows 7 64-bit SP1

Following recent reviews, we have added several new benchmarks to our testing, including SYSMark and Unity-based testing methodology. At present we have only tested a few CPUs on these new benchmarks.

Power Consumption

For our power consumption tests, we take the system as a whole with a GTX 770 installed and test at idle, long idle (monitor turned off by the OS) and OCCT (artificial CPU load).  We use the same PSU to keep efficiency on the same scale, although some of the data points are sub-20% efficienct for our power supply.  This means that arbitrarily the qualitative comparison is more important than the quantitative comparison.  Also of note is the Z97 motherboard we used for these tests implements an Adaptive voltage profile, meaning that artificial loads such as OCCT push the voltage higher than normal, increasing power consumption at load. Nevertheless the order of the results indicates a pattern from low power to high power.

Power Consumption - Long Idle  (GTX 770 Installed)

Power Consumption - Idle (GTX 770 Installed)

Power Consumption - OCCT  (GTX 770 Installed)

As expected, the 84W i7-4790 requires the most power, similar to the i7-4770K and i5-4690 that are also 84W.


Intel Core i7-4790, i5-4690 and i3-4360 Tested CPU Performance: Real World Benchmarks


View All Comments

  • rajod1 - Thursday, May 29, 2014 - link

    I agree, waste of cash to upgrade every generation. Old days were nice. Intel said they would hit 10 GHZ by like 2004 or something. LOL. So they hit a wall but still needed cash. Lots of suckers born every minute that will upgrade for 10 percent increase. Reply
  • Hrel - Monday, May 12, 2014 - link

    The fact that hyperthreading is disabled on the i5's on the desktop is infuriating. Enabling it costs them nothing. Agree on the 8 core at $300, it should still have HT though. Reply
  • Flunk - Monday, May 12, 2014 - link

    Definitely, turn on HT on i5 and give them all 4 cores and offer an 8 core i7 with HT. i3s can soldier on with only 2 cores if they really want to.

    A i7 5770 with 8 cores and HT is what they really need to bring out, but I think they're waiting for AMD to bring out something better first.
  • rajod1 - Friday, May 30, 2014 - link

    You may as well ask them to drop the I5 because if you put HT on a I5 its then I7. Reply
  • rajod1 - Friday, May 30, 2014 - link

    It does cost them something. They sell my I7s that way. Reply
  • bsim500 - Sunday, May 11, 2014 - link

    Thanks for this review - appreciate the effort that went into it as always.

    "Also of note is the Z97 motherboard we used for these tests implements an Adaptive voltage profile, meaning that artificial loads such as OCCT push the voltage higher than normal, increasing power consumption at load"

    So they've still got that dumb Haswell "feature" of stuffing the Vcore up by +0.1v when you least want it? Also why are your power consumption figures so high in general? My i5-3570 @ 4.0Ghz barely pulls 37w idle / 88w 4-threads prime with a 7870 discrete card (whole system (excluding monitor) measured at the wall). That's roughly 25w lower both idle & load than your i3 (with a 1600MHz idle vs 800Mhz Haswell's)! Just out of curiosity, what where the stock / default VID's like on the i3-4360 / i5-4690? ie, has the clock speed bump increased required voltages much? Thanks.
  • Ian Cutress - Sunday, May 11, 2014 - link

    Adaptive is an Intel specification implementation, but for some reason goes haywire with certain 100% load simulators (OCCT/AIDA).

    With regards power consumption seeming high:
    a) Low efficiency band of the PSU. Hence me stating qualitative analysis more relevant than quantitative. Need to keep the PSU consistent across all the tests, some tests require 2x/3x GPUs (e.g. X79). This is probably a large part of it, but all tests are therefore done on the same efficiency curve.
    b) Using a Corsair H80i with two fans and ODD plugged in. I move the USB devices to USB 2.0 so any USB 3.0 controller can power down, but it still all adds up.
    c) OCCT loading does the adaptive voltage thing, causing more power consumption at load from idle.

    We got ES chips to test, so retail might have adjusted slightly on the stock VID. Also VID can differ from chip to chip in the same bin, so it's not really a good measure. One CPU can have a high VID in the bin, while the next bin up we could get a low VID, and it all look a bit odd.
  • Daniel Egger - Sunday, May 11, 2014 - link

    IMNSHO it is really ridiculous to test all systems with the same outlandish special PSUs that no sane person would ever use. Why not have a testbed for single card systems with say a platinum 500W PSU (which should cover even the nasty R295X2 plus a Haswell K processor) and a separate one for crazy setups? With those far sub 20% loads even under full load it is nearly impossible to get useful readings not to mention comparable ones since at these low loads lots of funny effects kick in skewing the results... Reply
  • wetwareinterface - Sunday, May 11, 2014 - link

    The reason you don't have a low end psu on your test bench is it's a test bench.
    The one setup should handle anything you can possibly throw at it and then a little extra for good measure.

    Also having the exact same high wattage psu to test everything on eliminates the psu as a differentiating factor when testing multiple system configurations.

    and finally the sad truth is there are several reviewers working for anandtech each from home and each with whatever they have laying around to do said testing with...
  • Daniel Egger - Monday, May 12, 2014 - link

    > The reason you don't have a low end psu on your test bench is it's a test bench.
    The one setup should handle anything you can possibly throw at it and then a little extra for good measure.

    Exactly my point for suggesting a 500W PSU rather than something much lower that I would personally put into a build. That should be sufficient for any even just halfway reasonable setup.

    > Also having the exact same high wattage psu to test everything on eliminates the psu as a differentiating factor when testing multiple system configurations.

    Unfortunately that's not true. Very low output on high wattage PSUs skews the results quite a bit because they typically are not accurate enough when it comes to handling the low loads thus smearing over the results with their own losses. I assume this is also why our Greek friend here doesn't even bother to test loads below 5% (which would be 60W at a 1200W PSU, about twice as much as my current Haswell PC needs on an mostly Idle Windows desktop).

Log in

Don't have an account? Sign up now