Intel Skylake: Core i3 6100 review

This review was originally written before locked Intel chips based on the Skylake architecture - like this i3 - could be overclocked. Feeling that our review is a little out of date, we went back to the Core i3 6100 and pushed our particular chip to its limits, and the results are remarkable. In fact, it's possible to match base Core i5 6500 quad-core performance - so we thought we would update this review accordingly with all of our findings. However, do bear in mind that Intel is looking to close down i3 overclocking, so exploiting the current loophole may not be possible on future motherboard purchases.

Having dominated enthusiast gaming with its i5 and i7 line, it's fair to say that the baseline i3 line is often overlooked, but the reality is that in most games, Intel's dual-core, quad-thread processor is still capable of handing in creditable performance, even on the very latest titles that explicitly specify higher-spec CPUs for their recommended settings. Over the last year, our Core i3 4130 has served us well in our budget gaming build, but with the arrival of its new-gen Skylake successor, there are some genuinely impressive gains. Prices still need to settle down (we paid £93 for this chip - typically £10 over the odds for an i3) but the benchmarks and the gameplay performance are fascinating - and there is overclocking potential here that makes a great budget processor even more of a bargain.

We've already reviewed the key Skylake processors aimed at the higher-end enthusiast - the Core i5 6600K and the Core i7 6700K - finding them both to be best-in-class parts with impressive performance improvements over their predecessors in CPU-bound gaming scenarios. The boost to Skylake's raw capabilities comes from a number of factors - firstly that the processor features two generational improvements over its direct predecessor, Haswell (Broadwell only received a limited desktop release) and secondly that the prior DDR3 system RAM is replaced by DDR4, meaning higher levels of memory bandwidth. These factors are just as important for the dual-core Skylake i3, and in the case of the move to DDR4, possibly more so.

On top of that, the i3 also features all of the other architectural improvements enjoyed by the quads, including more internal PCI Express bandwidth, principally allowing for more, faster storage solutions to be attached. However, the chances are that the i3 will be paired with lower-end motherboards with more limited expansion opportunities - but as we'll shortly discover, there are performance benefits from choosing a better board to pair with the new dual-core CPU.

Our Core i3 6100 test kit

We bought the Core i3 6100 ourselves for £93. Other i3s are available - the i3 6300 adds 1MB of cache, increases the clocks to 3.8GHz, and mildly upclocks the integrated GPU, while the i3 6320 takes frequencies to 3.9GHz. We chose the bottom-end chip because it's the only one available for under £100.

For the purposes of our Skylake reviewing, MSI supplied us with its Z170A Gaming M5 motherboard, while Corsair supplied two sticks of 8GB of Corsair Vengeance LPX low profile DDR4, rated for 2666MHz. For testing the full extent of the overclock, we bought 3000MHz Corsair sticks, which overclocked with each to 3200MHz. Our stock 512GB Crucial MX100 SSD provided the storage for all those 50-gig test games, while power came from an existing Corsair HX750 PSU.

Clearly a mid-range board compared to the Asus Z170 Deluxe we tested with the i5-6600K, nonetheless, the M5 is a great product. SATA Express is included as standard, while two M.2 SSD slots are available as standard for ultra-fast PCI Express RAID - that's something the Asus board required a bundled PCIe riser board to achieve.

MSI produces more expensive, feature-rich Z170 boards, but the M5 has everything you need - the usual surfeit of USB 3.1 ports, plus the requisite Type C connector. Integrated video outputs consist of HDMI and DVI (more expensive boards tend to bundle DisplayPort), while the two main PCI Express slots have a metal protective sleeve to protect them from wear. The firm's tie-up with Killer continues, with the inclusion of its low latency LAN port plus accompanying software suite. Extra software supplied with the board includes a RAMdisk and the Xsplit livestreaming tool.

Our one gripe with the M5? No onboard power/reset buttons - a problem more for bench testers than real-life gaming perhaps, but a notable omission nonetheless. The M5's feature set is grounded in reality - indeed, our favourite feature is the back panel - the LAN port is illuminated very brightly, meaning that plugging in cables around the back in dark conditions (ie under virtually any desk) is a breeze. Really, why isn't this feature completely standard by now? In common with other Z170 boards, the Z170A Gaming M5 allows for RAM speeds over 2133MHz - crucial in our testing in this review.

Skylake's other notable enhancement - specifically its more granular overclocking capabilities - is now also available for locked chips like the i3. The K-series quads allow for both tweaking of the base clock and the multiplier when prior generations only really allowed for adjustment of the latter. At launch, base clock was almost completely locked down - we achieved a mere two per cent increase in performance (77MHz) before the system locked up. However, all major motherboard vendors have since released updated BIOSes for their Z170 boards, meaning that while the multiplier is still locked, the base clock is now tweakable. We could increase base clock on our chip to 120MHz, giving us a 4.44GHz overclock. However, even without overclocking the processor itself, pairing the stock i3 with faster DDR4 can produce substantial increases to performance.

But let's start by taking a quick look at some basic benchmarks, where we'll compare the new i3 with one of its Haswell predecessors - our trusty Core i3 4130 - along with the AMD challenger in this segment, the FX-6300. Additionally, we'll toss in our existing Skylake data from the quad-core end of the market to get an idea of the architecture's scalability across the range. The results are enlightening.

First up, it's worth pointing out that our Core i3 4130 operates with a clock-speed deficit compared to the new i3 6100 - it runs at 3.4GHz up against the 3.7GHz of the Skylake chip - much as we would have liked to run Haswell comparisons on a clock-for-clock basis, that just wasn't possible. Regardless, across our four tests, the new i3 is handing in performance that's around 17-26 per cent faster. The new chip may only have two cores and four threads, but its multi-threading performance is impressive - 3DMark physics is faster than the six-core FX-6300, both CineBench tests are only marginally slower, while only the x264 encoding tests sees the AMD chip command any kind of tangible improvement.

In fact, even comparisons with the FX-8350 and the Sandy Bridge era Core i5 2500K look surprisingly promising, especially in terms of the 3DMark physics test, which is the closest thing we have to any kind of gaming workload in what are predominantly synthetic and multimedia-orientated benchmarks. Overall, it's a solid start for a budget processor, and as we saw with the Core i5 6600K and the i7 6700K, basic tests like this are no match for putting Intel's technology to task with some challenging gaming content.

We set about benchmarking the new Core i3 using the same principles we used in assessing the Skylake Core i5 and the i7 - we seek to eliminate the GPU as a bottleneck by pairing the CPU with an overclocked Titan X running at 1080p, but we run the games at max settings (albeit with no multi-sampling anti-aliasing in most cases). The role of the CPU is to run game logic and prepare instructions for the GPU - compromising on quality settings would mean fewer elements to draw in any given scene and would not fully test the processor workload.

Additionally, we benched the i3 6100 twice, first of all using the full 2666MHz bandwidth of our Corsair Vengeance DDR4 modules, and then paring that back to 2133MHz in order to match the memory restrictions on the H170, B150 and H110 motherboards more likely to be utilised for budget builds. And as the benchmarks came in, the results were fascinating - in many CPU-bound scenarios, the i3 6100 is significantly faster with higher-speed RAM. We devised the gaming benchmark set-up to eliminate the GPU as the limiting factor in performance but what is clear is that not only is CPU pushed to the fore, memory bandwidth is too.

In the first set of benchmarks, we compare both of those i3 runs with the Skylake i5 and i7, in order to judge how the lower-end chip compares to its more expensive siblings. It's worth watching the video to get the complete breakdown of performance in context, because averages can be deceptive - on the face of it, as an average aggregate all across nine titles, the i3 6100 offers 80 per cent of the performance of the 6700K and 86 per cent of the raw throughput of the 6600K: not bad for an entry-level i3. However, while a strong argument can be made that an i5 offers a very similar real world experience to an i7, the same is not the case with the i3. Play Crysis 3, Ryse, Project Cars or Grand Theft Auto 5 to name just four titles, and the difference is often night and day. Our benchmark sequences - in common with everyone else's - can only offer one snapshot of any given game.

The two i3 runs are probably the more fascinating comparison in the table above. Consider the difference that 2666MHz memory makes to performance. That Ryse figure is no error - performance falls through the floor when running with lower levels of bandwidth, while faster RAM offers 11 per cent more performance on GTA 5 and Far Cry 4. And again, those figures are averages spread out across the benchmark run - it's noticeably higher at any given point during 'in the moment' gameplay. The Skylake platform standardises on 2133MHz RAM, and that's all that's available on cheaper boards - only the Z170 range officially lets you run memory faster, and it's clear that there are good advantages in doing so. Put simply, the memory controller in Skylake isn't saturated at the 2133MHz speed limit of the lower-end boards, meaning that gaming performance drops more than it should when you're in CPU-limited scenarios.

It's fair to say that there's a hierarchy of potential bottlenecks in any given PC gaming scenario. First of all, there's the graphics hardware, followed by the CPU. Our tests artificially propel the CPU to the fore, and by extension magnify the impact of limited memory bandwidth. The question is whether this extreme testing will be borne out in actual gameplay conditions, where you're more likely to be running with a lower-end GPU. That can be easily answered with this single screenshot. That's Crysis 3 on the i3 running with a relatively meagre GTX 950, and the game isn't even maxed - we're running on high settings, not the top-end very high. The CPU is fully tapped out in both scenarios, but we're also hitting memory bandwidth limits at 2133MHz, resulting in an 'in the moment' performance deficit of 6fps, along with higher degrees of stutter observed in the frame-times. It should be noted that this scene runs close to a locked 60fps with a Core i7 4790K, proof positive that as fast as the i3 is, there will be gaming scenarios where you really need a quad to get the job done.

How we measure performance

Our preferred form of measuring PC performance is to use the FCAT system pioneered by Nvidia. The idea here is remarkably straightforward - rather than use internal tools such as FRAPS to measure performance, FCAT does nothing more than apply a coloured border to the output of the host PC, with each individual frames marked up, ready for analysis.

The source PC is attached to an entirely separate computer, using a high-end capture card that acquires every frame produced. We've adapted our existing frame-rate video tools to work with the FCAT border mark-up system, allowing us - uniquely - to measure performance in context of what is actually being rendered. In short - abstract metrics like lowest frame-rate, highest frame-rate and stutter become much more meaningful when you can actually see what is causing them. We use repeatable scenes from our gaming suite to ensure as close to a like-for-like comparison as possible.

Our result tables are drawn from the FCAT metrics, but on top of that, the videos allow you to see exactly how each game performs: there, not only do you see frame-rate, but frame-time too - showing persistence of every single frame generated by the system. Riva Tuner Statistics Server also has its own FCAT support - we prefer to use Nvidia's version, but the advantage RTSS offers is the ability to gauge CPU and GPU load simultaneously - click on the shot above to see our Witcher 3 test sequence max out the new i5 6600K.

The bottom line is this - most reviews out there provide simplistic bar charts to gauge performance: great for an at-a-glance summary, but lacking in terms of understanding how hardware actually performs at any given point, and where the averaging effect serves to 'iron out' the in-the-moment performance differentials that matter most.

In short, the Core i3 6100 is a capable processor and almost the certainly the best in the sub-£100 price bracket, but if you're looking to get best performance and to better future-proof the platform, we'd recommend that you consider an entry-level Z170 board as opposed to cheaper alternatives. We are hearing rumours that some H170 boards may unofficially support memory overclocking too, which could save some money if true. On top of that, 2133MHz RAM is the absolute baseline - hunt around for 2666MHz sticks. They can be found at a reasonable price if you search - at the time of writing Ebuyer is selling 8GB of Kingston Hyper 2666MHz DDR4 for less than £45, easy enough to absorb into any budget build.

Moving on, the table above shows how the Skylake i3 compares to other budget processors. We're stacking it up primarily against the Core i3 4130, its predecessor from the Haswell launch, albeit running at a 300MHz deficit compared to its successor (ideally we would have had the later Core i3 4170 to test, but alas that was not possible). To make things fairer, we're also using faster RAM on the old Z97 platform, as well as standard 1600MHz sticks usable on the last-gen budget motherboards - in short, the Haswell equivalent to the 2133/2666MHz Skylake test. And finally, we're including AMD's FX-6300, although we could only bench that with 1600MHz memory - our Gigabyte AM3+ motherboard refused to run our 2400MHz DDR3 any faster without overclocking the processor itself.

The results show that the last-gen Haswell also benefits from faster RAM too. Ryse is once again crippled with slower memory and still plays like a dog even with faster memory, a situation that is even more impactful on the FX-6300. On the one had, it may well be the case that the CryEngine games are cache-limited, making their reliance on system RAM much more of an issue, but on the other, we saw similar issues in that game with a Core i5 2500K.

It's worth repeating that clock speeds are not like-for-like, but we are seeing improvements north of 20 per cent between Skylake and Haswell here, and it's actually the case that (CryEngine apart) a Core i3 6100 with 2666MHz DDR4 is generally on par or even a little faster than an older Core i5 2500K with 1333MHz DDR3 when both systems are paired with a GTX 970. The same set-up also sees Skylake beat the AMD FX-8350 (paired with 1600MHz DDR3) in every game we tested bar Crysis 3 and The Witcher 3. Of course, those chips beg to be overclocked in a way that the i3 never can, but the bottom line is that in many gaming scenarios, the new i3 is capable of performance that belies its dual-core status. In short: choose your components carefully - get your board choice right, buy the right RAM, and the Core i3 6100 forms the basis of a great, easily upgradable gaming platform. Choice to GPU is also important - at this performance level, AMD's driver overhead is an issue and we would recommend an Nvidia card.

All of which places the Core i3 6100 at the top of the table when it comes to sub-£100 processor performance, but the recent ability to overclock the previously locked Skylake CPUs can have an explosive effect on gaming performance - but do note that Intel is intent on closing down this loophole. With a base-clock tweak to 120MHz, and a voltage boost to 1.37v, we were solid at 4.44GHz. In our initial tests with the overclock, we restricted ourselves to the same 2666MHz DDR4, and it's safe to say that we were somewhat bewildered by the fact that some games saw tangible increases, but others saw only smaller increases to performance, not in line with a 20 per cent clock-speed boost.

How to overclock a locked Intel Skylake CPU

Since the days of the second-gen Core architecture - Sandy Bridge - and the iconic Core i5 2500K, overclocking has by and large been very simple. CPU clock-speed is defined by two factors: the base clock (aka BCLK, typically 100MHz) and the multiplier, which varies between processor. The Core i3 6100 has a 37x multiplier, giving 3.7GHz (100x37=3700).

The K series Intel chips allow users to increase the multiplier on the motherboard BIOS, resulting in higher CPU clock-speeds. The more you push the multiplier, the more voltage is required and the more heat generated. We don't recommend pushing voltage any higher than 1.3v.

The so-called BCLK overclock for non-K chips like the Core i3 6100 is different. The multiplier remains locked, but the base clock can be increased from the stock 100MHz. In the case of the i3, we were comfortably stable at 1.37v with a 120MHz BCLK. Factor in the 37x multiplier and we're at 4.44GHz. Voltage needs to be increased until the system is stable.

Overclocking via BCLK means that the entire system is sped-up - and that includes RAM. In the memory settings, you'll find that the RAM has also increased in speed, and it will need to be reduced to match the rating of the modules. Alternatively, you can try pushing RAM harder and increasing voltage there in order to achieve higher levels of memory bandwidth.

The BCLK overclock has some advantages, but it should be stressed that this is a hack of sorts, and not supported by every Z170 board. You should also be aware that the more voltage you push into the CPU, the hotter it runs. The i3 6100 runs perfectly well at stock speeds with the supplied cooler - pushing voltage higher could move the heat threshold beyond its capabilities, meaning that a third party cooler is required.

Our Core i5 6500 review testing - which included overclocking - revealed that getting the most out of an overclock required scaling up RAM bandwidth too, so it was not surprising to see that the same thing is true for the i3 too. We used the same Corsair Vengeance sticks rated at 3000MHz here, and we overclocked to 3200MHz with no effort whatsoever. This produced the kind of scaling in performance we expect, and the big surprise here is a Core i3 6100 at 4.44GHz, paired with 3200MHz RAM can marginally outperform a Core i5 6500 at stock speeds, when combined with base-speed 2133MHz memory. Of course, the i5 can also be overclocked and it can also be combined with faster RAM and obviously it pulls ahead in these scenarios.

However, there is some additional food for thought: we're currently working on a feature that sees just far the classic 2011 Core i5 2500K can be pushed in terms of CPU and RAM overclocking and whether it can still power the latest games effectively, and let's just say that initial results suggest that the near-legendary Sandy Bridge quad-core chip gets pushed very, very hard by our top-end dual-core i3 overclock results. And that's with the i5 running at 4.6GHz with 2133MHz DDR3. Still, for now, here's the Skylake comparison:

Intel Skylake: Core i3 6100 - the Digital Foundry verdict

There was some disappointment that we couldn't overclock the Core i3 6100 when we first received it for review, but regardless, even at stock speeds it proves to be a highly capable performer - particularly when paired with faster memory. The only problem here is that a Z170 board is required, and although prices for 'premium' speed RAM are stabilising, you'll still be paying more for higher bandwidth modules. In both cases, we would suggest not skimping on this area of the build. A non-Z170 board can be upgraded to quad cores i5s and i7s, but what's clear is that you'll still be held back in terms of future-proofing without access to higher-speed memory and CPU overclocking.

In terms of our test results, Skylake's domination is self-evident, but we did make one pleasant surprise - AMD's cheaper FX-6300 managed to put up a decent challenge against the last-gen i3, which isn't bad bearing in mind it's a £78 processor that's just celebrated its third birthday - ancient in technological terms. However, generally speaking, with or without the benefit of faster RAM and overclocking, the Core i3 6100 is considerably faster, the surrounding chipset is light-years ahead in terms of features and upgrade paths, and with DDR4 prices dropping hard, upgrading to new-gen RAM isn't the budget deal-breaker it once was.

The Skylake Core i3 6100 is an accomplished product bearing in mind its price-tag, and in conclusion, it's worth emphasising just how important performance is at this end of the market. High-end games are becoming more CPU-intensive, making it much more likely that the processor will be the bottleneck during any given gaming session - especially so at the budget end of the market. Every cycle matters in the more demanding areas of many modern titles, and in our testing, the Core i3 6100 is best in class - and now it's overclockable, a great chip just got even better.

Will you support the Digital Foundry team?

Digital Foundry specialises in technical analysis of gaming hardware and software, using state-of-the-art capture systems and bespoke software to show you how well games and hardware run, visualising precisely what they're capable of. In order to show you what 4K gaming actually looks like we needed to build our own platform to supply high quality 4K video for offline viewing. So we did.

Our videos are multi-gigabyte files and we've chosen a high quality provider to ensure fast downloads. However, that bandwidth isn't free and so we charge a small monthly subscription fee of £4.50. We think it's a small price to pay for unlimited access to top-tier quality encodes of our content. Thank you.

Find out more about the benefits of our Patreon

ncG1vNJzZmivp6x7psHRqJ6apZWne6%2Bx02ibop%2BZqa6tss6upZ2qqWJ%2FcX2UZqCnrJWheqS70Z5komtda35xfIyrnK%2Bhlaw%3D