Overclockers Australia Forums

OCAU News - Wiki - QuickLinks - Pix - Sponsors  

Go Back   Overclockers Australia Forums > General Topics > Overclocking & Hardware


Sign up for a free OCAU account and this ad will go away!
Search our forums with Google:
Thread Tools
Old 29th April 2016, 1:01 PM   #1
headin2001 Thread Starter
headin2001's Avatar
Join Date: Apr 2007
Location: Sunshine Coast
Posts: 2,947
Default Asus Strix R7 370 DirectCU II OC 4GB - User Review

Asus Strix R7 370 DirectCU II OC 4GB - User Review


Iím baaaaack! Today I will be showing you the Asus Strix R7 370. The card is based on the now aging AMD Pitcarn architecture, originally featured in the Radeon HD 7870 and 7850 graphics cards. The tuned architecture has been able to hold itself well over the years, and AMD has worked hard to continue to develop and improve over the years, through refinements in its inner workings, but also through increased clock speeds thanks to the reduction of the node process to 28nm.

Features are well in line with the current range of video cards on the market. The card itself features support for DirectX12, Mantle, OpenGL 4.5, OpenCL 2.0 and AMDís GCN 1.0 Stream processors.

Asus has created its own custom designed PCB layout and power delivery system featuring their 5-phase Alloy Super Power II VRM. Also a standout is the Strix DirectCU II cooling, with two (2) 8mm heat pipes transferring the heat away, and its 0dB fan system for silent operation when the GPU temperature is below 65įC.


Features / Software


I mentioned this back in my GTX 960 Strix Review (found here : http://forums.overclockers.com.au/sh....php?t=1188878 ), so those who read through, you can probably skip over this sections.

Asus have taken things a little further with a complete ground-up design based around Auto-Extreme Technology. This is a first for the industry. The one thing I find that stands out with Asus products is the persistence in being market leaders in not only design and performance, but also its manufacturing process.

Basically, Asus has developed an automated system for the production of its products and have removed the Ďhuman errorí factor and increased quality control of the products. In the process they have saved the workers from all those harsh gases released with the use of flux. Donít look at this the wrong way, itís a win-win for all. We, the end users can purchase a product of a very high quality standard, and for the workers in the factory, they no longer will have their lives shortened from breathing the deadly fumes every day.

The final product is a nice neat looking card with fantastic quality control and a more reliable unit. All pluses in my books.

Here is a video to explain the process better.


GPU Tweak II Software

Included with all Asus Graphics cards is their own GPU Tweak software. I mentioned it in my previous review of the Asus GTX 960 Strix 2GB (Link - http://forums.overclockers.com.au/sh....php?t=1188878)
Initial setup was quick and simple. The layout is very easy to navigate and customise to suit what you want to see/use. The setup menu allows for keyboard shortcuts, and also to decide on start-up options. Another option is to unlock further settings past the limited clock, power, and memory settings for the more advanced users looking to push things further than most.

Factory profiles are included for:
Silent operation (a slightly under-clocked setting)

Gaming (which is the default clock speeds)

OC (a minor overclocked setting)

Or User Profile, which can be setup and saved to suit yourself, or you can add extras to suit your own personal needs.

Photos / Comments

The product comes well packaged in a cardboard box, with sufficient packing around to ensure there is little chance for it to be damaged.

Inside the box we have the graphics card itself, a Crossfire bridge, driver CD, manual and a nifty little Strix sticker.

Moving onto the cooler, we can see from the top that Asus has a nice Owl theme going, with the menacing eyes looking back at you. The cooler is setup to only turn on when the card reaches the pre-set temperature. The fans measure 75mm. 2 x 8mm copper heat pipes transfer the heat from the core to the aluminium fin array. There is also what appears to be a solid finned heat sink to cover and spread the heat of the VRM componentry.

The power delivery for the card is supplied through a single 6-pin connector, located on the rear edge. Asus have used their Super Alloy Power II component for the power delivery, featuring a 5-phase setup for cleaner power supply and longevity. Above the 6-pin power connector, there are red indicator LEDís which change to green when the power is connected correctly. Nice little touch there. It can also aid when troubleshooting, should there be a problem. The connector is also rotated 180į from the usual, with a little notch in the PCB. This makes removal of the power a lot easier than other cards.

Asus recommends a minimum of 500W power supply with a minimum of 24 Amps on the 12V rail.

On the backside of the card, we can see how physically short the PCB actually is. The heatsink radiator and fans from the DirectCU II cooler actually protrude past the end of the card. My guess is to aid with the airflow through the cooler, which is quite effective. Overall, the card measures 21.5cm in length and 12.5cm in width.

The card can support up to six (6) monitors through the rear connections using AMD Eyefinity. It includes four (4) different connectors: 1 x HDMI, 2x DVI and 1 x Display Port.

headin2001 is online now   Reply With Quote

Join OCAU to remove this ad!
Old 29th April 2016, 1:01 PM   #2
headin2001 Thread Starter
headin2001's Avatar
Join Date: Apr 2007
Location: Sunshine Coast
Posts: 2,947

Test Setup

My GPU test rig uses an Intel i7 6700K overclocked to 4.5 GHz on both core and cache, with the Vcore set to 1.35V. Being a modern processor with up to 8 threads available, the chance of a CPU bottleneck on this system is minimal. The CPU is paired up with a set of G.Skill Ripjaws V 3466C16 RAM in a 4 x 4GB configuration, with the timings set to XMP with all timings locked to reduce the chance of change between reboots, and all connected to an Asus ROG Maximus VIII Hero motherboard (see my review here).

Power comes from a Silverstone Strider Gold Evolution 1200W, to ensure there is no starvation of available power. Cooling for the CPU is taken care of by a custom loop, consisting of an EK Supremacy Evo CPU block, D5 pump housed in a XSPC dual bay reservoir, Alphacool XT45 420 3 x 140mm radiator with BitFenix 140mm Spectre Pro PWM fans set to low. The GPU’s are installed within a Corsair Obsidian 750D, with the side of the case removed, more for access and to allow clean air for repeated benchmark runs.

Most of my results are from the use of built-in benchmarks, and for those that do not, I have come up with some manually played sections which I have recorded the results with the use of FRAPS. The benchmarks are repeated 3 times to ensure repeatability and provide an average for the figures provided for Minimum and Average frames per second. For all benchmarks excluding Crysis 3, Windows 10 Pro 64bit is used. Windows 7 Pro x64 is used for Crysis 3 due to compatibility issues.

For Nvidia graphics cards, I have used the in-built Dynamic Super Resolution (DSR) for 1440p benchmarking due to being limited to a 1080p monitor. For AMD graphics cards, the Virtual Super Resolution (VSR) feature was enabled in order to test the higher resolutions.

All overclocked results are based on three (3) passes of Unigine Heaven Xtreme, and overclocked results are shown within the 1080p result graphs. The software used for overclocking and monitoring is Asus GPU Tweak II.

Test System
• CPU: Intel i7-6700K at 4.5GHz (45 x 100MHz)
• Motherboard: Asus ROG Maximus VIII Hero
• Memory: 4 x 4GB G.Skill Ripjaws V 3466C16 (XMP Profile)
• PSU: Silverstone Strider Gold Evolution 1200W
• SSD: Samsung 840 250GB
• Case: Corsair Obsidian 750D
• CPU cooler: Custom water loop
• Operating system: Windows 10 Pro 64-bit / Windows 7 Pro 64-bit

Nvidia Graphics Cards
• EVGA GeForce GTX 780 ti – 1204MHz Boost, 8GHz GDDR5 (GeForce 358.50)
(Note, this card is pre-overclocked due to a custom bios, no memory timings have been adjusted, just base clock and boost removed)
• Asus GeForce GTX 960 Strix - 1254MHz GPU (1317MHz boost), 7.2GHz GDDR5 (GeForce 355.82)
• Asus GeForce GTX 750 ti OC – 1072MHz GPU (1150MHz Boost), 5.4GHz GDDR5 (GeForce 358.50)

AMD Graphics Cards
• Asus Radeon R7 370 Strix 4GB - 1,050MHz GPU, 5.6GHz GDDR5 (Catalyst 15.8 driver)


Games tested:
Battlefield 4
Fallout 4
Grand Thief Auto V
Tomb Raider
Middle-Earth - Shadow of Mordor
Alien: Isolation
Metro 2033 Redux
Crysis 3

Battlefield 4

Battlefield 4 is my go-to game and I literally book time during the week to play this. For this, I use the start-up sequence for Shanghai. It’s full of DOF changes, lots of lighting, water, texture rich, facial close-ups… The works really. It’s a timed run using FRAPS for 110 seconds, from the moment you walk through the door to the loading bay until the RHIB drops into the water.

Fallout 4

Fallout 4 is lacking a built-in benchmark. A little disappointing, but at least we can get an actual run through to give a more accurate feel of the game. For this, I have a run through near the beginning of the game, with sections which have little overhead to show the highest frames achievable, to lots of forest and water, to groups of people and some nuclear fallout.

It was also necessary to make changes to some .ini files to disable v-sync, something that is not easy to change in-game, and every change to resolution or settings results in v-sync being enabled again. I had to double check every time that it was not affecting the results. So beware! FRAPS is used to record the data.

Grand Thief Auto V

GTA V results are taken from the 5th built-in benchmark test. As the game engine does not have any pre-set graphic profiles, I have come up with a custom setting to stress the GPU’s but also try to keep it repeatable between the differing architectures. As the game engine runs with varied population and vehicles in the city, there can be a fair difference between runs, so this was run 3 times with an average result listed. GPU memory use is a big one on this title, so the more you have, the more eye candy you can adopt.

Settings are:
FXAA – on
MSAA, Vsync off,
Texture quality - very high
Shader quality – high
Shadow and reflection quality – normal
Reflection MSAA – off
Water Quality – high
Particle and grass quality – very high
Soft shadows – soft
Post FX – ultra
In-game DOF effects – on
Anisotropic filtering – X16
Ambient occlusion – high
Tessellation – very high

Tomb Raider
Fantastic game; lots of foliage, hair effects, cloth, texture – it has it all. The benchmarks are run on the Ultra pre-set on 1080p and 1440p utilizing the built-in benchmark for repeatability.

Middle-Earth: Shadow of Mordor

Quite a taxing game, especially at higher resolutions and definitely brings all of the cards to their knees. As I only have access to a 1080p monitor, I have taken advantage of Nvidia’s DSR (Dynamic Super Resolution) and AMD’s VSR (Virtual Super Resolution) in order to test higher resolutions. This worked out to be 2880 x 1620, so the performance shown will be tougher than the 1440p standard. The Ultra pre-set was used. It appears that 2GB cards seem to suffer at this setting, so the 3GB+ cards have a little breathing room.

Alien: Isolation

A game that runs just like the movie of old. Full of suspense and impending death, this game can really get the heart going in places. There are lots of lighting effects, and tessellation to really tax the graphics settings. Alien: Isolation is getting a little old now, but it is still a good title to push modern graphics card to the limits and provide scope for comparison.

For this title, I again utilised the built-in benchmark. The score represents the results from the 5 separate tests involved while using the Ultra Graphics pre-set.

Metro 2033 Redux

Metro is a nice little Post-Apocalyptic FPS. Lots of environmental effects, lighting, and physics with a good story behind it all. Basically like the original, but tweaked using the setup from Last Light.
The Benchmark is run using the built-in tool. Quality setting on very high, with no SSAA, AF set to 4X, normal motion blur, very high Tessellation, and with Vsync and PhysX set to off. Each graphics setting is run 3 times, and the average recorded in the graph below.

Crysis 3

Can it run Crysis?

Well, here we find out. Crysis 3 is still one of the most graphically intensive and taxing games around, and will bring even the top tier cards down to show their worth. Here, I run the game at very high settings, with a run through based on one of the checkpoints on the way to the Railroad, in the Welcome to the Jungle chapter. All results are recorded using Fraps, with the average of three (3) runs recorded.


Software used:
Futuremark 3DMark11
Futuremark 3DMark Firestrike
Futuremark 3DMark Firestrike Extreme
LuxMark v3.1 – Hotel (Complex Scene)
Unigine Heaven – Basic
Unigine Heaven – Extreme
Unigine Valley – Extreme HD
Cinebench R15 - OpenGL


Power measurements are taken as the peak load power recorded during Unigine Heaven Extreme. This provides a good deal of power usage to cover most situations. Some games may require extra power over this, but it is a very good baseline.

Due to the nature of electronics, the usage of power is greatly dependant on CPU type, voltage and speed, number of system fans, hard drives, the PSU itself, and other connected peripherals. This measurement is purely a guide for what this particular test system achieves. Always refer to the manufacturer’s guides for the minimum PSU requirements to suit the GPU.

Sound Measurements are taken using an iPhone app, by Skypa. My room in quiet times flattens at 55dB with just the main system in idle, all case fans on low, and with the phone located 30cm away from the case position. The base system itself is barely audible, with only the usual fan hum and D5 pump noise.

Temperature measurements are via the Asus software suite, GPU Tweak II. Measurement was taken before and after 2 runs of Unigine Heaven Xtreme Benchmark for maximum heat soak. The side of the computer case was left off for the duration of the tests, with an average room temperature of 28įC. All fan setting were left on the software’s Default Auto Settings.


As mentioned earlier, Asus bundle their very own GPU Tweak II software with their gaming cards. I utilised this software as well as Unigine Heaven in order to find what I would class as a safe 24/7 overclock. Generally, this meant leaving Unigine Heaven in the Extreme pre-set to loop for 2 hours and not experience any artefacts or driver crashes.

The Asus Strix R7 370 already comes with a decent factory overclock, from the AMD base clock of 975MHz on the core clock to a nice 1050MHz operating frequency – a good 75MHz there. On top of this, I was able to manage an extra 140MHz, making a total of 215MHz overclock from stock. Fairly strong in my opinion.

For the memory side of things, the GDDR5 deployed is rated at 5600MHz from the factory, left unchanged from the reference models. During my testing, I noticed in 3DMark11 that my scores when overclocking the RAM decreased as I increased the speed. All I can attribute this to is there is a set of memory straps employed within the bios. Testing identified 5600-5700MHz was the speed which provided the best possible scores whilst in the synthetic benchmarks, with 5700MHz being the best. Running a speed higher than 5700MHz, even by 10MHz, would result in the score decreasing. And not just a minor score differential, but by a good 100-200 points! The plus side though, for those who want the highest MHz possible, is that I was able to take it to a stable 800mhz overclock without issue, although the scores were slightly lower than the 5700MHz optimum.

Here is a GPUz screenshot for verification.

Last edited by headin2001; 6th May 2016 at 4:31 PM.
headin2001 is online now   Reply With Quote
Old 29th April 2016, 4:10 PM   #3
headin2001 Thread Starter
headin2001's Avatar
Join Date: Apr 2007
Location: Sunshine Coast
Posts: 2,947


Generally, I found the card to be quite capable in most titles, albeit, not particularly with Ultra settings applied. For me, I found high to be the happy medium for the card to shine in my testing. I love the eye candy myself, but high on most modern titles is quite enough for my eyes.

The ability to easily crossfire the R7 370 is a good plus for those looking to upgrade on a budget, and its relative power use isn’t too much for a decent power supply to handle. The inclusion of 4GB of ram I find also a great benefit, with some games memory usage pushing over the more common 2GB range of old.

Cooling wise, Asus has done a fantastic job. Although I haven’t tested a reference based card, its ability to not only run with a decent factory overclock but also still run at an acceptable temperature and minimal noise is a testament to how hard Asus works to make this a clear point of difference. The 2 x 75mm fans are a perfect size for it, and do a very good job to cool the aluminium heat sink below. Plus, the shroud is just cool too. I love the 0dB noise at idle and general browsing too. It not only helps reduce power usage, but that noise is eliminated at the same time.

Now, onto a couple of negatives. First up, the lack of a back plate. I know, it’s a budget mid-range card, and for a small card, it’s not particularly necessary. But the look of it and small expense it would have overall on the card, I feel would be worth it. Not that the back of the card is ugly, or full of pins to cut your fingers on which is testament of the outcome achieved by the Auto-Extreme Technology manufacturing process.

Secondly, Firestrike. You may have noticed in my graphs I have omitted an overclocked score for this GPU. Well, I have a very good reason for that. During my testing, I noticed that during any test in the current 3DMark, including Firestrike, Firestike Extreme, Cloudgate or Sky Diver, I had an unusual case of the GPU down-clocking to 925MHz, regardless of what I had set in the software. I tried a multitude of different drivers, a clean Windows 7 and Windows 10 install, all with no sign of rectifying the issue. A quick Google search led me to believe that some manufacturers’ cards have the same issue, where others don’t. The issue is unknown, but, on the plus side there is no problem present in any of the other tests I conducted. I think it is particularly important to note that this issue doesn’t arise during actual gaming and therefore is most likely an issue associated with this particular suit of 3DMark synthetic tests.

The Asus R7 370 Strix can be found at most local computer stores, with a price in the range of $225 to $275 AUD. (search http://www.staticice.com.au/cgi-bin/...wadres=1&pos=2)

If you enjoyed this review of the Asus R7 370 Strix, be sure to check out my review of the Asus GTX 960 Strix found here: ***link***

I hope you enjoyed my review. Please feel free to comment/question in the thread below. Fingers crossed for another to follow in the next week.

Last edited by headin2001; 29th April 2016 at 4:15 PM.
headin2001 is online now   Reply With Quote
Old 29th April 2016, 6:02 PM   #4
Join Date: Jan 2009
Location: Nulkaba
Posts: 1,775

Another excellent review
mad_mic3 is offline   Reply With Quote
Old 29th April 2016, 6:06 PM   #5
Paulie_AU's Avatar
Join Date: Nov 2013
Location: Brisbane
Posts: 3,408

Nice review mate.

Be interesting to see how crossfire would go vs say a 380x or even a 390. Might have the grunt then to fully need the 4gb.

Paulie_AU is online now   Reply With Quote
Old 29th April 2016, 6:43 PM   #6
headin2001 Thread Starter
headin2001's Avatar
Join Date: Apr 2007
Location: Sunshine Coast
Posts: 2,947

Thanks guys. I actually really rated this card and think I could find a place for it in my arsenal.

Next review is of the big card. Some touch ups to do on it yet, but 90% there. These are so time consuming, I really have a lot of respect for the guys that do this professionally and with tighter timeframes.

I have a cooler review to come too, so keep an eye out.
headin2001 is online now   Reply With Quote
Old 30th April 2016, 5:01 AM   #7
Sgt Bilko
Sgt Bilko's Avatar
Join Date: Mar 2015
Location: Canberra(ish) 2581
Posts: 705

Very nice review mate!

The 370 is a weird card for me, from a gaming aspect is is perfectly suited for 1080p high settings with a nice Freesync monitor, problem is because it's GCN 1.0 it doesn't support it (Freesync that is) :/

I'm very much looking forward to the next ones you've got lined up though
AMD Community Manager

Comments are my own unless stated otherwise.

Last edited by Sgt Bilko; 30th April 2016 at 5:03 AM. Reason: I Derped
Sgt Bilko is offline   Reply With Quote
Old 30th April 2016, 10:45 AM   #8
ginger_nuts's Avatar
Join Date: Sep 2011
Location: Morphett Vale, SA
Posts: 1,164

So this card would make a nice HTPC card for gaming @1080.

If you still have the card, can i bother you to please measure the height from the top of the pcie bracket to the heat pipe ?

My Silverstone case is extremely limiting on height.
The world will keep turning
HWBot Profile = Ginger_Nuts81 Member of Australia OC team.
ginger_nuts is offline   Reply With Quote
Old 2nd May 2016, 11:43 PM   #9
Join Date: Apr 2015
Location: Qld
Posts: 115

I do agree the lack of a backplate is a little annoying, but the card performs alright for it's cost. Another $100 will get you a R9 380 or GTX960 which is a good performance boost of course.
Phanteks Evolv mATX: i7-4790K/GTX980Ti
Successful trades: Over $1000+
jji7skyline is offline   Reply With Quote
Old 7th May 2016, 4:22 PM   #10
headin2001 Thread Starter
headin2001's Avatar
Join Date: Apr 2007
Location: Sunshine Coast
Posts: 2,947

Originally Posted by ginger_nuts View Post
So this card would make a nice HTPC card for gaming @1080.

If you still have the card, can i bother you to please measure the height from the top of the pcie bracket to the heat pipe ?

My Silverstone case is extremely limiting on height.
For sure, its a good little card for such a task. I had no problem running pretty much every game I tried at 1080p and reasonable quality settings.

Think the measurement your chasing measures around 18-20mm. Message me if you need a photo to check.
headin2001 is online now   Reply With Quote


Sign up for a free OCAU account and this ad will go away!

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT +10. The time now is 9:32 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
OCAU is not responsible for the content of individual messages posted by others.
Other content copyright Overclockers Australia.
OCAU is hosted by Micron21!