ASRock Phantom Gaming X Radeon RX580 8G OC Review: A Solid Rookie Effort

AMD's Radeon RX 580 is old news at this point. But in a way, that makes it the perfect card for ASRock's debut as a graphics company. The platform is stable. The competition is established. And the stakes are relatively low. If you're not already familiar with the GPU at the heart of this Phantom Gaming X Radeon RX580 8G OC, check out our AMD Radeon RX 580 8GB Review. Or, go even further back in time to our AMD Radeon RX 480 8GB Review, when the Ellesmere GPU made its grand entrance. Today's introduction takes those two years of valuable lessons and applies them to a mainstream card that was clearly designed with cost savings in mind.

The Radeon RX 580's performance is already well-known, so this card's success depends on whether its relatively small thermal solution, lack of a backplate, and absence of eye-catching LEDs can push pricing low enough to undercut competing 580s. Since we got our hands on the Phantom Gaming X before it landed on store shelves, we can only guess that ASRock wants to go up against Sapphire's Nitro+, Gigabyte's Gaming 8G, and PowerColor's Red Dragon, the least-expensive Radeon RX 580s out there. Unfortunately, U.S.-based readers may have to wait a while to find out: ASRock is only selling into South American and APEC countries to start.

Specifications

GPU (Codename)Ellesmere
Shader Units2304
Base & Boost Clocks1380 MHz
Memory Size & Type8GB GDDR5
Memory Clock2000 MHz
Memory Bandwidth256 GB/s
Fans(2) 85mm
Ports(1) DVI-D, (1) HDMI 2.0, (3) DisplayPort 1.4
Power Connectors(1) 8-pin PCIe
Dimensions26.7 x 10.5 x 3.5cm
Weight598g
Warranty1 Year

Unboxing, Look, And Feel

A weight of just 598g tells us that ASRock went with a conservatively-sized thermal solution to cool AMD's Ellesmere GPU. Still, measuring 26.7cm from the slot bracket to the end of the fan shroud, this is a fairly long graphics card. The 10.5cm height and 3.5cm width measurements keep ASRock's Phantom Gaming X within the dimensions of a true dual-slot form factor.

Two 8.5cm fans sit in 8.7cm openings. A total of nine rotor blades per fan are optimized for moving air through the heat sink, so they generate more static pressure than fans designed for turbulent flow.

ASRock shaves off some cost by not using a backplate. That's a sensible decision in our opinion; it wouldn't have done much for cooling, and it isn't needed for stabilization since the thermal solution is so light.

Peering in from the bottom, we can see that ASRock uses horizontally-oriented fins. This is our preference because it allows some hot air to exhaust from the slot bracket. The alternative, fins rotated to face vertically, pushes hot air down toward the motherboard and against the side of your case.

An eight-pin auxiliary power connector visible from the top is rotated by 180 degrees, making it easier to access. On a pricier model, you might expect ASRock's logo to be back-lit with LEDs. That's not the case for its Phantom Gaming X though, and we're fine with this.

The open back side allows warm air to exhaust into your case. Moreover, the use of a short PCB means ASRock's heat sink overhangs the board a bit.

The slot bracket features five familiar outputs. In addition to the single DVI connector, you get one HDMI 2.0 port (particularly useful for VR HMDs) and three DisplayPort 1.4-ready interfaces. Ventilation holes cut into the plate allow some hot air to travel through the horizontal fins and out of your chassis.

The following screenshot from GPU-Z conveys this card's maximum clock rates. In reality, the Phantom Gaming X's power and temperature limits mean those frequencies often aren't sustainable, though.

Test System & Measurement Methods

We introduced our new test system and methodology in How We Test Graphics Cards. If you'd like more detail about our general approach, check that piece out. We've upgraded the CPU and the cooling system since then to make sure that nothing's holding back graphics cards as fast as this one.

The hardware used in our lab includes:

Test System
Hardware
Intel Core i7-6900K @ 4.3 GHz
MSI X99S XPower Gaming Titanium
Corsair Vengeance DDR4-3200
1x 1TB Toshiba OCZ RD400 (M.2, System SSD)
2x 960GB Toshiba OCZ TR150 (Storage, Images)
be quiet Dark Power Pro 11, 850W PSU
Windows 10 Pro (All Updates)
Cooling
Alphacool Eisblock XPX
5x be quiet! Silent Wings 3 PWM (Closed Case Simulation)
Thermal Grizzly Kryonaut (Used when Switching Coolers)
Case
Lian Li PC-T70 with Extension Kit and Mods
Configurations: Open Benchtable, Closed Case
MonitorEizo EV3237-BK
Power Consumption MeasurementContact-free DC Measurement at PCIe Slot (Using a Riser Card)
Contact-free DC Measurement at External Auxiliary Power Supply Cable
Direct Voltage Measurement at Power Supply
2 x Rohde & Schwarz HMO 3054, 500 MHz Digital Multi-Channel Oscilloscope with Storage Function
4 x Rohde & Schwarz HZO50 Current Probe (1mA - 30A, 100 kHz, DC)
4 x Rohde & Schwarz HZ355 (10:1 Probes, 500 MHz)
1 x Rohde & Schwarz HMC 8012 Digital Multimeter with Storage Function

Thermal Measurement
1 x Optris PI640 80 Hz Infrared Camera + PI Connect
Real-Time Infrared Monitoring and Recording

Noise Measurement
NTI Audio M2211 (with Calibration File, Low Cut at 50Hz)
Steinberg UR12 (with Phantom Power for Microphones)
Creative X7, Smaart v.7
Custom-Made Proprietary Measurement Chamber, 3.5 x 1.8 x 2.2m (L x D x H)
Perpendicular to Center of Noise Source(s), Measurement Distance of 50cm
Noise Level in dB(A) (Slow), Real-time Frequency Analyzer (RTA)
Graphical Frequency Spectrum of Noise

MORE: Best Graphics Cards

MORE: Desktop GPU Performance Hierarchy Table

MORE: All Graphics Content

Create a new thread in the Reviews comments forum about this subject
This thread is closed for comments
19 comments
Comment from the forums
    Your comment
  • Ulikedat
    At least it's a looker and the competition just got a bit more heated (no pun intended).
  • alextheblue
    I have to wonder, since aiming for a cut-down budget design, why not an RX570? An entry level RX580 with halfway decent cooling can be had new for $270. The cheapest 570 I've seen is a reference model for $250. Push a budget aftermarket RX570 closer to $200 and undercut everyone else. Clocks and TDP would be lower (especially if they stuck close to reference), which would have further reduced board and cooler costs.

    Or perhaps yields are so good at this point, that there just aren't many cut-down Ellesmere chips getting pushed out the door?
  • BulkZerker
    @alextheblue they are pushing the bang for buck at every point it seems. It begs to wonder if some "cheap" (like as5) thermal paste would help or not. Certainly an aftermarket heatsink would help the temps.
  • TJ Hooker
    If this card is like most 580s, it will respond quite well to lowering the core voltage. With some undervolting (and maybe a slight underclock), you could likely improve power, thermals and noise noticeably with no (or little) impact on performance.
  • AnimeMania
    What's going on with the numbers for GTX 1060 3GB with certain games like Hitman and DOOM, they can't really be that bad, can they?
  • gca.roberts
    The 1R0 marking on the GDDR5 VRM inductor means 1.0uH, not 1.0mH.
  • Sleepy_Hollowed
    This is quite a nice card, especially for those looking to switch to FreeSync and the open drivers that AMD provides (For accelerating stuff like data compression or video encoding).

    Like all cards, I just wish it was available, this crypto craziness is on the downslope for now, but you never know when cards are just going to be missing from the shelves for months.
  • Olle P
    Anonymous said:
    I have to wonder, since aiming for a cut-down budget design, why not an RX570? ...
    Around here there's almost no price difference between a 570 (4GB) and 580 (8GB). The latter is significantly better at the all important 1080p so that's where to make profit.

    Anonymous said:
    If this card is like most 580s, it will respond quite well to lowering the core voltage. ...
    One can only hope. The 180W drawn is a bit steep IMO.

    Anonymous said:
    What's going on with the numbers for GTX 1060 3GB with certain games like Hitman and DOOM, they can't really be that bad, can they?
    Seems off topic...
    The 3GB is a cut down version of the 6GB, with fewer ROPs and less memory bandwidth. 3GB VRAM is also insufficient to run newer games efficiently.
  • AnimeMania
    Anonymous said:
    Anonymous said:
    What's going on with the numbers for GTX 1060 3GB with certain games like Hitman and DOOM, they can't really be that bad, can they?

    Seems off topic...
    The 3GB is a cut down version of the 6GB, with fewer ROPs and less memory bandwidth. 3GB VRAM is also insufficient to run newer games efficiently.

    According to another review on Tom's Hardware, the GTX 1060 3GB had 68.1 FPS on Hitman at Ultra Levels.

    https://www.tomshardware.com/reviews/nvidia-geforce-gtx-1060-graphics-card-roundup,4724-2.html

    The GTX 1060 3GB in this article had 21.2 FPS on Hitman at Very High Levels. I don't think it is off topic if I am questioning the reliability of the benchmarking process used here. I was just wondering why the results seem to fluctuate so wildly. One value is 3 times higher than the other.
  • alextheblue
    Anonymous said:
    Around here there's almost no price difference between a 570 (4GB) and 580 (8GB). The latter is significantly better at the all important 1080p so that's where to make profit.

    That was my point. There's less competition for cheap RX 570s. Their 580 design cuts down on costs across the board (pun intended). They could cut down even further with the lower-TDP (and presumably cheaper) RX 570 and undercut the entire field substantially. An RX 570 at ~$200 would be enticing for budget builds.

    When they first came out there was often a substantial price difference between full Ellesmere and cut-down Ellesmere. That's why I'm speculating that there just isn't enough supply of 570 chips to make this possible.
  • alextheblue
    Anonymous said:
    The GTX 1060 3GB in this article had 21.2 FPS on Hitman at Very High Levels. I don't think it is off topic if I am questioning the reliability of the benchmarking process used here. I was just wondering why the results seem to fluctuate so wildly. One value is 3 times higher than the other.
    What you're looking at is very likely the result of drivers over the last ~15 months. If it was testing methodology, why doesn't the 6GB version suffer in Doom and Hitman? What it looks like to me is that the current drivers aren't being as efficient with memory allocation and you're dipping into system RAM on the 3GB model, or there's some bug that's causing memory usage issues. If the card was performing poorly in all games I'd say there's an issue with the card. But it performs "as expected" for some titles, and not for others. But Doom and Hitman? Drivers are what I'd blame.

    Also, look at both those articles again. Look at The Witcher 3 1440p average FPS for the 1060 6GB. The fastest 1060 scored 52.1 on the older article, and the unnamed (probably more pedestrian to go against the el cheapo ASRock 580) 1060 6GB on the newer article scored 52.2. So there are driver improvements, but it seems they unintentionally caused problems with certain games for the 3GB model. You can probably drop settings a notch and get RAM consumption back under control and performance would rocket back up.
  • ohenryy
    It's about time Asrock has it's own VGA lineup. Most others have it and they are actually the ones I was most looking for to see doing some graphics cards.
    It's a good start with the 580, the card is a well worth competitor to the 1060's out there and I do think it's a better choice. If you have a Freesync, this card is a no brainer.
    Now we just need to start seeing the card available and at normal prices...
  • Olle P
    Anonymous said:
    ... Their 580 design cuts down on costs across the board (pun intended). They could cut down even further with the lower-TDP (and presumably cheaper) RX 570 and undercut the entire field substantially. ...
    I'm not so sure about that.
    1. The tested RX 580 doesn't seem to have that substantial reductions in production cost. It still has 8GB of fast memory and four outputs.
    2. I don't think that currently the wholesale price of 570 and 580 GPU chips doesn't differ a lot from each other.
    3. Even if ASRock were able to sell 570 cards at a significantly lower price the vendors could just increase their own margins instead.
  • Olle P
    Anonymous said:
    The GTX 1060 3GB in this article had 21.2 FPS on Hitman at Very High Levels. I don't think it is off topic...
    Sorry about my "off topic" remark. I didn't check the gaming results very close and definitely not in comparison to other tests.
    Anonymous said:
    What you're looking at is very likely the result of drivers over the last ~15 months. ...
    Updated graphics drivers is a possible explanation, yes.
    You can for example see a similar impact on Rise of the Tomb Raider, where the "average" FPS has dropped from 69 to 44 with the same settings.

    Other possible factors:
    * Different CPU and RAM speed/latency used for the tests.
    * Impact from Spectre/Meltdown patch. (Should affect the faster cards even more though.)
    * Different specs between the previously tested MSI Gaming X and the card now used. (The MSI card is ~10% faster than reference while running in "OC mode".)
    * Updated (more demanding) game files.
    * Ran the test using a different part of the game?
  • alextheblue
    Anonymous said:
    1. The tested RX 580 doesn't seem to have that substantial reductions in production cost. It still has 8GB of fast memory and four outputs.
    2. I don't think that currently the wholesale price of 570 and 580 GPU chips doesn't differ a lot from each other.
    3. Even if ASRock were able to sell 570 cards at a significantly lower price the vendors could just increase their own margins instead.
    1) They cut wherever they could, aside from wanting to compete head-to-head with 8GB models. They designed a cheaper, smaller board design with reduced componentry costs. This would be even easier with a 570-specific design, given lower power demand (cheaper PD circuitry) and lower TDP (even smaller/cheaper HSF). A 570 at a low price point could also slash VRAM in half to 4GB since it would be pitted against other 4GB 570s and 1060 3GB.
    2) I have no data on this. Link? That's why I was speculating about 570 chip supply from AMD.
    3) The vendor margins bit is true of anything they release. They could try to sell this cheaper 8GB RX 580 and get it bumped by $30-50 (better than $200+ over MSRP like it was!). As supply loosens and mining ETH winds down, that will start to rectify itself, because there are lots of vendors. All you need is good supply and one vendor selling at/near MSRP to bring pricing in-line at the others.

    At least until something takes over for ETH and is also ASIC-resistant (for a while). :/

    Anonymous said:
    * Impact from Spectre/Meltdown patch.
    Hah, I almost forgot about that one! My only Intel systems currently are non-gaming machines so I tend to not think about it. It wouldn't explain the huge crash in performance between the 3GB and 6GB models though... but it does offset some of the performance gains they made in their drivers.
  • Olle P
    Anonymous said:
    Anonymous said:
    2. I don't think that currently the wholesale price of 570 and 580 GPU chips doesn't differ a lot from each other.
    2) I have no data on this. Link? That's why I was speculating about 570 chip supply from AMD.
    ... All you need is good supply.
    Pricing is a speculation from my side, based on the supply argument below.
    Since the retail prices have gone way up one can expect all parties involved wanting to get a piece of the action and doing so by increasing their sale prices.
    Then there's the question of what constitutes a "good" supply. There's no apparent shortage of AMD based graphics cards in Sweden today, at the current prices, so if vendors could make a profit by cutting the consumer prices and sell more they probably would. If the vendors has paid (way) above MSRP to get the cards in the first place they can't reduce the prices without making a loss on each sale.
  • alextheblue
    Anonymous said:
    If the vendors has paid (way) above MSRP to get the cards in the first place they can't reduce the prices without making a loss on each sale.

    That's true, and it slows down overall price reductions across the market. But a card which is new to the channel and sold at a lower price wouldn't have that problem. The problem would again become supply-related retailer gouging. If they could supply an infinite number of a new card at a low price, all it takes is one vendor sticking to MSRP to keep the rest from gouging.
  • HugoMeira
    What happened to GTX 1060 3GB? He simply died in 3 games that are dependent on vram at very high or ultra. This board has always been a nvidia cheat. I prefer RX470 / 570 or RX 580 4GB.
  • HugoMeira
    What happened with GTX 1060 3GB? Simply died in 3 games that are dependent on vram at very high or ultra. This board has always been a nvidia cheat. I prefer RX470/570 or RX 580 4GB.