Choosing the parts
In order to find out the real world potential of the G4560 and how it stacks up with the erstwhile dominators of the budget circuit, we decided to choose hardware in accordance with the budget segment we are tackling. After all, we need to make sure that the parts we zero in on are what not beyond the capabilities of the buyer of the CPU being tested. Considering the price tag of the G4560, we decided to use the following parts:
redirected here Motherboard : Asus EX – H110M – V (mATX)
Considering that the G4560 has no overclocking capabilities (one of the main draws for the Z170 chipset), and is best used in the budget gaming sector rather than the business sector (where the H170 chipset dominates), an H110M board sounds perfect.
see here RAM : Kingston Value 2133Mhz 2x4GB kit
It was a toss-up going to 8GB of RAM over 16GB, but price friendliness and the linear cost per gigabyte of today’s RAM sticks made the extra 8GB useless considering we are dealing with games, not memory intensive programs. It is also worth mentioning that the One-DIMM-per-channel restriction of the H110M platform contributed to the decision (aside from this being the budget sector that is). We decided to pick the first kit we could find lying around, and this was it.
http://haciendaparrillaalta.com/producto/vino-tinto-dulce-hacienda-parrilla-alta/?add-to-cart=475 Power Supply Unit : Coolermaster MasterWatt 600 Lite (80Plus White)
Juicing this bench is the 600W Masterwatt Lite power supply, with an 80Plus White efficiency rating. Considering the range of GPUs we were choosing, we decided to go for a beefier PSU than what would normally go with this kind of setup. Since we are dealing with a locked SKU, and at max a 1060, 450W is sufficient for all intents and purposes.
Asus GTX 1050 Expedition
Asus GTX 1050Ti Expedition
Asus Strix Gaming GTX 1060
We chose the following GPUs with the target resolution in mind (1080p) along with the cost of the rest of the parts. While the 1050 might not be ideal for said scenario, it happens to be one of the cheapest GPUs available at the time of testing (April 2017), and can deliver a serviceable frame rate with lower graphical presets.
Monitor : AOC 2476W
Entry level 1080p@60Hz monitor with support for DVI and HDMI.
The tests were carried out on an open bench and consisted of the stock cooler that comes with the G4560, mounted on a layer of Coolermaster Mastergel Maker Thermal Paste. We used a clean install of the stock Windows 10 Creator Update, with Nvidia Driver 378.78 (verified to be stable at the time of testing).
Overwatch is Blizzard’s first attempt at a new IP in years, fusing MOBA-based characters with FPS and team-based elements (also known as a “team shooter”, which is akin to, and popularized by, Team Fortress 2). The game is not GPU intensive, but it does make use of advanced shadow and reflection techniques, along with render scaling, that can impact FPS significantly. Usually anything above 48 FPS is considered playable, with 60 FPS recommended, however, for competitive players, getting close to 120 FPS or above is a huge bonus, as high refresh rate gaming has taken off substantially in the ESports arena.
We used Player vs AI matches to gauge real world performance, choosing to bench FPS all the way from our first spawn to the POTG. Frametime graphs of the matches were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows. The top 3 graphic presets (High, Ultra, Epic) were tested across all 3 GPUs listed, and data listed below.
As can be seen in the above Data, the G4560 holds up just fine, with pretty decent lows and average framerates almost always exceeding the recommended 60 FPS, although those targeting high refresh rate gaming (100+ FPS) should look towards investing in a good GPU, or a better CPU if they wish to break into the 144Hz+ segment, consequently needing to increase the budget of the overall build as well.
Grand Theft Auto V
GTA: V has been out for a while now, and Rockstar made sure it stayed relevant for years to come. Its complex PC settings allowed great performance optimization for all systems requiring fine tuning. As there are no existing presets (barring the ones that can be manually applied via configs, a method that we will not use), we need to normalize and split it up between common setting levels available, aka Normal, High, and Very High. (Very High meaning we use the maximum settings permissible). We did not touch Advanced Options, and limited MSAA to 2X, with AF at 4X for Normal and High.
Once again, 60 FPS is the recommended framerate for this game, although 48-50 cuts it just fine. We opted for a standard mission to benchmark, it being the 2nd mission with Franklin and the 3rd mission in the game. We breezed through the travelling sections of the mission, and during the shootout, shot both the car fuel trail and the overhead gas tank to include explosions in the benchmark. Once again, frametime graphs of the entire mission were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
From the above data, we can conclude that while the lows provided by the cards do not meet the permissible framerate criteria of 48 FPS, they are not that far from the mark. This ensures that the G4560, when paired with any of the above cards, and the right preset, should be able to output smooth gameplay without any issues.
Counter Strike: Global Offensive
CS: GO signified the start of the new E-sports era back when it released in 2012 after the immense success of Counter Strike: Condition Zero/Source/1.6. CS: GO mostly leverages the CPU, so we will be seeing minimal differences in the stats across GPUs, mostly seeing benefits in the % lows provided.
For benching, we used a Workshop mod, FPS benchmark, which follows a predefined trail in a map filled with 100+ NPCs, nested smokes, and flashbangs, to determine the system’s capabilities. (You can thank the nested smokes for the extremely bad % lows BTW). Once again, frametime graphs of the benchmark runs were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
As can be observed above, the G4560 is more than enough for gaming at high refresh rates, and will be perfect for setting up machines in cafes/tourneys, although it is incapable of going past the 180 FPS threshold, something that might be a turn off for those who wish to go 240Hz, or who wish to stream their gameplay (although that’s a really tall expectation from this kind of setup).
Deux Ex: Mankind Divided
Horrible ports have plagued the industry since time immemorial, more so in the past years. However Studio Nixxes managed to do a good job for Square Enix’s Rise of the Tomb Raider, and did it again with Deus Ex: Mankind Divided. It is built on the Dawn Engine, the same one used by Hitman: Absolution.
The ingame benchmark serves as a good representation of actual gameplay, and consists of 90 seconds of a single pan across multiple character models. We benched said scene at 3 different presets, namely, High, Very High, and Ultra. Owing to the issues DX12 currently has with Nvidia drivers (such as frametime variances), we avoided that variable entirely and chose to stick with DX11. Once again, frametime graphs of the preset runs were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
It is here that we start to see the G4560 slightly lacking, with only the 1060 managing to push it past the 48 FPS at all, and the rest bordering on unplayable territory. However I wouldn’t read too much into it, as Deus Ex: Mankind Divided has been shown to be very demanding in the first place, some even going as far as to call Ultra beyond the reach of non-enthusiasts.
Tom Clancy’s Rainbow Six Siege
Unsurprisingly, Rainbow Six Siege is still a very popular game among the masses, for its return to the roots of “tactics triumphs over run-n-gunning”, with some MOBA class based aspects added in for good measure and extra replayability. Siege comes with a benchmark included which serves as a good indication of how a system will handle actual matches, which lasts for barely 40-50 seconds. (Although one could argue that other maps are much more taxing than the one used in the benchmark, it doesn’t sound like a really pressing issue that could affect our findings, so we decided against it).
We ran the benchmark at the High, Very High and Ultra presets, and documented the results. Once again, frametime graphs of the matches were generated through FRAPS and data analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
The above data illustrates that the G4560 will do just fine with any of the cards above, easily clearing the 60 FPS requirement, and in fact capping out at 120 FPS with more powerful GPUs, making this a good fit for high refresh rate gaming. This is in part due to the game being heavily GPU bound.
The Witcher 3: Wild Hunt
After a rather worrying launch (owing in part to bad drivers at the time), CD Project Red managed to fix most of Witcher 3’s problems with a patch, leaving a masterpiece that leveraged both CPU and GPU to their limits (ala GTA V, though not as complex). And while Witcher 3’s graphic options make no sense to a gamer as much as they do to photographers, they can be easier to sift through via the presets distributed between 2 tabs, GRAPHICS and PREPROCESSING.
For consistency, we will not alter PREPROCESSING and leave it at Medium preset, while varying the Graphics Presets between Medium, High, and Ultra. We would definitely recommend disabling Nvidia Hairworks if you are facing issues breaking into playable FPS territory on lowend systems. For the benchmark, we started it off right after the Golem fight in “Lilacs and Gooseberries”, where we tried out Qwen, Arden and Igni, got on our horse, and raced Vesemir to the next checkpoint, all the while drinking in the beautiful scenery. Frametime graphs of the runs were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
It is interesting to note that although the VRAM consumption of Witcher 3 is 1800 MB, it still struggles with the 1050 (2GB VRAM), barely able to output playable gameplay. We observed good scaling among the 3 presets, signifying good optimization on CD Project RED’s part. It is also worth noting that there is virtually no difference going from 1050Ti to 1060.
Metro Last Light
Developer 4A Games and Publisher Deep Silver went the extra mile to remaster their original titles, Metro Last Light and Metro 2033, giving us a combined package named Metro Redux, which consisted of various improvements and tweaks to update the game to modern standards, including the use of a new engine, addition of melee attacks, extended level sections, and better AI, to name a few.
Metro Last Light’s included benchmark program leverages both CPU and GPU equally, and is regarded as a standard for the average gaming machine, even today. We ran said benchmark with SSAA and PhysX disabled, Normal Motion Blur, AF 16X and Very High Tesselation, while varying Texture Quality from Medium to High, and finally, Very High. Frametime graphs of the preset runs were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
As observed, Metro is able to comfortably make it past the 60 FPS threshold, with most % lows doing just fine. It is also worth noting that the linear improvements observed are in part due to amazing optimization on 4A Games’ part, and highlight the benefits of pairing a better GPU with the G4560.
Battlefield 1 is one of the latest games to utilize core count along with core frequency, making it one of the more futureproof titles out there. For reasons mentioned in the Deus Ex Benchmark, we will not be enabling DX12. For the benchmark, we played through the 1st mission of “Through Mud and Blood”, beginning when we first assume control of Black Bess, all the way up to when it stalls in deep mud.
We ran benches using the Medium, High and Ultra presets, all the while ensuring that we followed the same path per run, to guarantee uniformity in data. Frametime graphs of the preset runs were generated through FRAPS and analysed via FRAFS Bench Viewer to calculate avg FPS, 1% lows and 0.1% lows across all 3 GPUs.
As can be expected, the G4560 is unable to scale linearly, as is expected considering it is simply a dual core, although its high clock speeds edge it into playable territory at higher presets.