Supermicro C9Z490-PGW Motherboard Review: Comet Lake with 32 CPU PCIe Lanes

More widely known for its server-grade models, Supermicro always launches a small number of consumer motherboards, sometimes with some extra flair and hardware we don’t see from the regular vendors. This time around, Supermicro’s top-tier C9Z490-PGW uses a PLX chip which enables the board to operate with dual PCIe 3.0 x16 or quadruple PCIe 3.0 x8 slots. This is combined with dual PCIe 3.0 x4 M.2 slots from the chipset, a 10 gigabit Ethernet controller, and a Wi-Fi 6 interface which makes the C9Z490-PGW versatile for a wide array of users. 

Supermicro is one of the most recognizable brands in the server and workstation market. Still, as we saw in our review of the C9Z390-PGW, Supermicro is consistently injecting its ‘server’ grade DNA into its desktop models. The difference between Z390 and Z490 isn’t as stark as it could be, with the main attribute coming in the way of networking support, with an integrated Wi-Fi 6 MAC, which allows users to utilize CNVi modules. For Intel’s launch of the 10th generation Comet Lake processors, Supermicro unveiled a pair of Z490 models, the C9Z490-PG and C9Z490-PGW, with the only difference being that the PGW comes with a Wi-Fi 6 interface, while the PG does not.

The Supermicro C9Z490-PGW is one of the most unique Z490 models for several reasons, with a combination of unique and server inspired aesthetics, as well as an interestingly premium feature set. It represents its SuperO series of motherboards, which offers server-grade quality and a standard consumer-focused model. Touching briefly on the design of the C9Z490-PGW, it uses a blend of black and silvers to create a classy two-tone theme, with solid black aluminum power delivery heatsinks and SuperO metal reinforcement on both the PCIe and memory slots. The C9Z490-PGW drops integrated RGB LED lighting and RGB headers and doesn’t play on the lack of RGB support as a marketable feature like some vendors do.

Obviously the big feature is the PLX switch, enabling 32 PCIe 3.0 lanes on this motherboard. The use of PLX switches on mainstream motherboards was rife in the time of the Z77 platform, however it has fallen by the wayside, mostly due to the increased cost as the company that used to make these switches was acquired, and the price was risen to a more consumer-unfriendly price point. With this switch, the board can support two add-in cards at a full x16/x16, or four cards at x8/x8/x8/x8, all at PCIe 3.0 speeds (because Comet Lake is PCIe 3.0 and this switch is PCIe 3.0 only). This opens up a number of avenues for users wanting to enable, for example, a Comet Lake-based storage system with RAID cards. To get this many lanes would otherwise require a different platform, usually in the high-end desktop space, or a Xeon. On top of the PCIe lanes, there are also dual PCIe 3.0 x4 M.2 slots, with four available SATA ports supporting RAID 0, 1, 5, and 10 arrays and networking through a 10 gigabit Ethernet controller, Wi-Fi 6 with additional support for BT 5.1 devices. It also includes a premium onboard HD audio codec, with lots of USB support onboard, including a stacked rear panel and plenty of USB headers located around the edge of the PCB. It also has a modest level of memory support, with capability for DDR4-4000, with a capacity for up to 128 GB.

In our performance testing, we saw the expected levels we would associate with a board that is running Intel’s default power settings. Supermicro boards often run at strict Intel defaults, whereas consumer motherboards are more liberal with Intel’s suggestions for power limits and turbo levels. If this is taken into account, the C9Z490-PGW performed competitively, especially against the ASUS ROG Maximus XII Hero WiFi when comparing it when running without ASUS’s enhancements. These had no real impact on gaming performance, but it performed slightly lower than other Z490 models, some of which have mult-core enhancement features enabled by default. Our system tests showed that power consumption is noticeably higher than other models on test, predictably down to the PLX chip. It also has longer POST times than other Z490 models, which is a common theme for Supermicro boards due to a focus on professional-level elements. Out of the box, default DPC latency performance wasn’t too great, but it is still an acceptable score.


The Supermicro C9Z490-PGW undergoing thermal VRM testing

Overclocking with the C9Z490-PGW wasn’t as straight forward as first would seem. It uses a competent 8+2 phase power delivery, but the firmware is restricting the capabilities. The only way to see a noticeable uplift in performance is to manually adjust the PL1 and PL2 power limits within the BIOS. Without making these adjustments, we saw no real benefit to overclocking our Core i7-10700K, even when overclocking to 5.1 GHz. We saw thermal throttling at 5.2 GHz, and unfortunately, the board’s VDroop control is quite poor, with a much higher load CPU V-Core than is set in the BIOS. This caused havoc in our power consumption testing while overclocking. In our power delivery thermal testing, we also noticed that the VRMs got quite warm, and the CPU area of the socket was much hotter than it should be, especially for an ATX model. While this board is overclocking capable, it’s designed to fly much more at stock than it is overclocked. Given the specialized features, we’re ok with that.

The Supermicro C9Z490-PGW was initially launched with an MSRP of $395, but it is currently available to purchase at Newegg for just $360. This puts it directly against models such as the ASRock Z490 Taichi ($370), the GIGABYTE Z490 Aorus Master ($389), and the ASUS ROG Maximus XII Hero ($399), but the Supermicro is set apart with the use of a PLX chip. The Supermicro C9Z490-PGW is a different sort of motherboard, but still with a marketing strategy focused on the gaming market, but not conforming to include some of the highly marketable ‘gaming’ features such as RGB. It’s a solid board with a solid feature set, but SuperO is not as widely known as other gaming brands such as Aorus or ROG. 

Read on for our extended analysis.

Supermicro is more commonly known for its professional workstation and server-grade motherboards, and much of this can be seen across the C9Z490-PGW. First of all, there is no integrated RGB LED lighting, nor does it include any RGB headers. Instead, the SuperO branded Z490 model uses black heatsinks on a matte black PCB, with the silver SuperO armor on the full-length PCIe 3.0 and memory slots giving a nice contrast. The SuperO logo within the plastic rear panel cover illuminates, which glows white, which puts this as one of the least flashy premium Z490 models on the market. That being said, Supermicro instead relies on its feature set, and users looking for a board with more ‘pizazz’ surely won’t have to look far.

The most prominent aspect of the Supermicro C9Z490-PGW is that it has four full-length PCIe 3.0 slots. This is because it is using a Broadcom PEX8747 PLX chip. The PLX chip essentially muxes the CPU lanes, allowing for up to 32 lanes across the four full-length slots. These can operate at either x16/x0/x16/x0 or x8/x8/x8/x8. One of the main benefits of muxing PCIe lanes is it offers more expansion capabilities with RAID controllers, FPGA PCIe add-on cards, compute cards, as well as additional network controllers. Located in the middle of each pair of the full-length slots is a single PCIe 3.0 x1 slot.


The Broadcom PEX8747 PLX chip on the C9Z490-PGW

For storage there are a pair of PCIe 3.0 x4/SATA M.2 slots, with the top slot supporting up to M.2 2280 drives and the second slot allowing for up to M.2 22110 drives. Despite the Z490 chipset supporting up to six native SATA ports, the Supermicro uses just four of these. The four SATA ports include support for RAID 0, 1, 5, and 10 arrays. Memory support is limited to DDR4-4000 officially, which is considered low compared with other Z490 models. The four memory slots support a maximum capacity of up to 128 GB and include SuperO metal armor reinforcement, with support for dual-channel memory.

In the bottom right-hand corner is a two-digit LED debugger, which can help diagnose POST issues. This is also coupled with an internal speaker, which is common on Supermicro boards, which beep upon POSTing and specific beep codes assigned to different issues. The C9Z490-PGW is also using an unconventional front panel header orientation for the power and reset jumpers, with a right-angled design instead of straight angled. This should help with cable management. The board uses six 4-pin headers for cooling, with two designated for CPU fans, one for a water pump, and three for chassis fans. 

Looking at the power delivery, it is using a 10-phase design, which is driven by an Infineon XDPE12284C PWM controller operating in an 8+2 configuration. It uses premium Infineon TDA21490 90 A power stages for the CPU, with slightly lower spec Infineon TDA21535 power stages for the SoC. This power delivery is quite capable, with a maximum output of 720 A for the CPU, although it uses just one 8-pin 12 V ATX power connector. Despite this, it is more than capable of pushing an Intel Core i9-10900K to its limits on paper.

Keeping the power delivery cool is a pair of heatsinks, which aren’t connected by a heat pipe like other premium models. Both heatsinks feature an aluminum top with a large surface area with fins to direct airflow, but they don’t have much overall mass. This means that the power delivery is heavily reliant on good passive airflow within a chassis.

The C9Z490-PGW uses a Realtek ALC1220 HD audio codec to power both the rear panel audio connections and the front panel audio header. This is assisted by two large audio capacitors and four small ones, with another shorter and stubbier one just above these. The audio area is separated from the rest of the board’s controllers, although the HD audio codec isn’t protected by an EMI shield.

The rear panel is a single USB 3.2 G2x2 Type-C, one USB 3.2 G2 Type-C, two USB 3.2 G2 Type-A, and two USB 3.2 G1 Type-A. The networking has two RJ45 ports, with one powered by an Aquantia AQC107 10 GbE controller, with another by an Intel I219-V Gigabit PHY. The W in PGW stands for Wi-Fi, and as such, it includes an Intel AX201 Wi-Fi 6 interface, which also includes support for BT 5.1 devices. For audio, it includes five 3.5 mm audio jacks and S/PDIF optical output powered by a Realtek ALC1220 HD audio codec. Users looking to use Intel’s integrated graphics can use the DisplayPort 1.4 and HDMI 2.0a video output pairing. Finishing off the rear panel is a small but handy clear CMOS button.

What’s in The Box

The accessories bundle includes four SATA cables, a quick reference guide, a rear I/O panel, and two antennas for the Intel AX201 Wi-Fi 6 adapter. Also present is an M.2 installation screw kit, a sticker sheet for marking up cables, and a driver and software installation disc.

  • Quick reference guide
  • Driver/software installation disc
  • Rear I/O shield
  • 4 x SATA cables
  • Cable sticker sheet
  • Front panel header adapter
  • 2 x Intel AX201 antennae
  • M.2 installation screw pack

Like other vendors, Supermicro hasn’t changed much in its firmware design since the jump from Z390 to Z490. Using a very similar UEFI BIOS to the previous C9Z390-PGW, the C9Z490-PGW only differs in the menus’ complexity, with a new set of options for Comet Lake’s Thermal Velocity Boost (TVB) feature. The SuperO firmware uses a uniform GUI, which includes a primarily black color scheme, with shades of grey and blue throughout. The text is white, with highlights in a darker shade of blue to signify which option is currently selected. The firmware also has two primary modes, including an EZ mode and Advanced, including all of the customizable options and functions.

When booting into the BIOS for the first time, users will be greeted with the EZ mode, including a variety of key and vital information. From top to bottom, the EZ mode includes information about the BIOS version installed onto the 256 Mb BIOS chip and information about the installed processor, memory, and storage. Users can enable X.M.P 2.0 profiles on supported memory at the click of a button, with a fan profile information panel that resembles a tachometer. On the right-hand side, users can choose between three profiles specific overclocking profiles, including default, OC mode, and auto-tuning, which affords the board the ability to overclock the processor based on set parameters within the firmware.

Some of the board’s most important settings for maximizing performance come under the TDP Configuration settings. This is where users can configure the power limit settings, which are inherently set to Intel default specifications, with no profile or settings to enable a form ofmulti-core enhancement (MCE). This means that when a processor is installed into the C9Z490-PGW, it runs to Intel’s recommendations without any tampering to PL or Tau, which most vendors implement on its firmware by default to give it an edge over the competition. Although the board is more than capable of handling extra power and overclocks, the decision is with the end-user.

Gallery: Supermicro C9Z490-PGW BIOS Gallery

The rest of the board’s firmware includes overclocking options, which can be found under the aptly named overclocking tab. This includes multiple settings, including CPU frequency configuration either by all-core or per-core, and options for BCLK control, voltage controls for multiple areas, including CPU VCore, CPU PLL, VSCIO, as well as advanced memory options for altering things such as frequency and latencies. Overall the firmware is neatly presented, includes a fan profile customization utility, and includes multiple security options, which the vast majority of Supermicro motherboards include. Despite this being a consumer Z490 motherboard, Supermicro plays to its strengths. The board, which resembles the firmware, shows the board is more than capable of use in a professional environment. 

The only piece of software included with the C9Z490-PGW is the SuperOBooster utility. This is also supported by the Realtek Audio HD Manager software, which accompanies the drivers for the ALC1220 HD audio codec. This allows users to customize and tweak the audio settings and add reminiscent effects of the Windows XP days. 

It includes a mixture of various software elements, including overclocking functionality for both the CPU and memory and making on-the-fly voltage and load-line calibration adjustments. One interesting element to consider in our testing is that when we made alterations within the software, upon rebooting and entering the firmware, these changes would be consistent, meaning that both the software and firmware work hand in hand. Users can also customize the fan profile settings under the Thermal tab, with customization options available per header, not including the water pump header, and sync them up. The last tab allows users to update the firmware with the latest version available to download from the Supermicro servers.

Gallery: Supermicro C9Z490-PGW Software Gallery

The Supermicro C9Z490-PGW is an ATX motherboard with a premium controller set, with the inclusion of a Broadcom PEX8747 PLX chip. The PLX chip allows for muxing which means the four full-length PCIe 3.0 slots which can operate at x16/x0/x16/x0 or x8/x8/x8/x8, with the 16 lanes from the CPU essentially doubled up (peak bandwidth is still limited to 16x upstream). It also includes a PCIe 3.0 x1 slot, as well as a pair of PCIe 3.0/SATA M.2 slots and four SATA ports with support for RAID 0, 1, 5, and 10 arrays. The C9Z490-PGW can officially accommodate DDR4-4000 UDIMM memory, with a maximum capacity of up to 128 GB supported across four memory slots. For cooling, the board includes six 4-pin headers, with two for CPU fans, three for chassis fans, and one dedicated to water pumps.

Supermicro C9Z490-PGW ATX Motherboard
Warranty Period 3 Years
Product Page Link
Price $395
Size ATX
CPU Interface LGA1200
Chipset Intel Z490
Memory Slots (DDR4) Four DDR4
Supporting 128 GB
Dual-Channel
Up to DDR4-4000
Video Outputs 1 x HDMI 2.0a
1 x DisplayPort 1.4
Network Connectivity Aquantia AQC107 10 GbE
Intel I129-V GbE
Intel AX200 Wi-Fi 6
Onboard Audio Realtek ALC1220
PCIe Slots for Graphics (from CPU) 4 x PCIe 3.0 (x16/x0/x16/x0, x8/x8/x8/x8) (PLX)
PCIe Slots for Other (from PCH) 1 x PCIe 3.0 x1
Onboard SATA Four, RAID 0/1/5/10 (Z490)
Onboard M.2 2 x PCIe 3.0 x4/SATA
USB 3.1 (20 Gbps) 1 x USB Type-C (Rear panel)
USB 3.1 (10 Gbps) 2 x USB Type-A (Rear panel)
1 x USB Type-C (Header)
USB 3.0 (5 Gbps) 2 x USB Type-A (Rear panel)
2 x USB Type-A (One header)
USB 2.0 4 x USB Type-A (Two headers)
Power Connectors 1 x 24-pin Motherboard
1 x 8-pin CPU
Fan Headers 2 x 4-pin CPU
1 x Water Pump
3 x 4-pin Chassis
IO Panel 2 x Antenna Ports (Intel AX201)
1 x HDMI 2.0a output
1 x DisplayPort 1.4 output
2 x USB 3.2 G2 Type-A
1 x USB 3.2 G2 Type-C
2 x USB 3.2 G1 Type-A
1 x RJ45 (Aquantia)
1 x RJ45 (Intel)
1 x Clear CMOS button
5 x 3.5 mm audio jacks (Realtek)
1 x S/PDIF Optical output (Realtek)

The rear panel for a premium Z490 model is one of the most scarce for USB we have seen, with one USB 3.2 G2x2 Type-C, two USB 3.2 G2 Type-A, and two USB 3.2 G1 Type-A ports. Users can add another single USB 3.2 G2 Type-C port, two USB 3.2 G1 Type-A, and four USB 2.0 ports through the use of internal headers. It includes two video outputs consisting of a DisplayPort 1.4 and HDMI 2.0a output, with five 3.5 mm audio jacks and S/PDIF optical output powered by a Realtek ALC1220 HD audio codec. The C9Z490-PGW includes an Intel AX201 Wi-Fi 6 interface, which is the only difference between the slightly cheaper C9Z490-PG model. Wired networking includes a premium Aquantia AQC107 10 GbE controller, as well as an Intel I219-V Gigabit PHY. 

Test Bed

As per our testing policy, we take a high-end CPU suitable for the motherboard that was released during the socket’s initial launch and equip the system with a suitable amount of memory running at the processor maximum supported frequency. This is also typically run at JEDEC subtimings where possible. It is noted that some users are not keen on this policy, stating that sometimes the maximum supported frequency is quite low, or faster memory is available at a similar price, or that the JEDEC speeds can be prohibitive for performance. While these comments make sense, ultimately very few users apply memory profiles (either XMP or other) as they require interaction with the BIOS, and most users will fall back on JEDEC supported speeds – this includes home users as well as industry who might want to shave off a cent or two from the cost or stay within the margins set by the manufacturer. Where possible, we will extend out testing to include faster memory modules either at the same time as the review or a later date.

Test Setup
Processor Intel Core i7-10700K, 125 W, $374
8 Cores, 16 Threads 3.8 GHz (5.1 GHz Turbo)
Motherboard Supermicro C9Z490-PGW (BIOS 1.1)
Cooling ID-Cooling Auraflow X 240mm AIO
Power Supply Corsair HX850 80Plus Platinum 850 W
Memory G.Skill TridentZ DDR4-2933 CL 14-14-14-34 2T (2 x 8 GB)
Video Card MSI GTX 1080 (1178/1279 Boost)
Hard Drive Crucial MX300 1TB
Case Corsair Crystal 680X
Operating System Windows 10 1909 inc. Spectre/Meltdown Patches

Readers of our motherboard review section will have noted the trend in modern motherboards to implement a form of MultiCore Enhancement / Acceleration / Turbo (read our report here) on their motherboards. This does several things, including better benchmark results at stock settings (not entirely needed if overclocking is an end-user goal) at the expense of heat and temperature. It also gives, in essence, an automatic overclock which may be against what the user wants. Our testing methodology is ‘out-of-the-box’, with the latest public BIOS installed and XMP enabled, and thus subject to the whims of this feature. It is ultimately up to the motherboard manufacturer to take this risk – and manufacturers taking risks in the setup is something they do on every product (think C-state settings, USB priority, DPC Latency / monitoring priority, overriding memory sub-timings at JEDEC). Processor speed change is part of that risk, and ultimately if no overclocking is planned, some motherboards will affect how fast that shiny new processor goes and can be an important factor in the system build.

Hardware Providers for CPU and Motherboard Reviews
Sapphire RX 460 Nitro MSI GTX 1080 Gaming X OC Crucial MX200 +
MX500 SSDs
Corsair AX860i +
AX1200i PSUs
G.Skill RipjawsV,
SniperX, FlareX
Crucial Ballistix
DDR4
Silverstone
Coolers
Noctua
Coolers

Not all motherboards are created equal. On the face of it, they should all perform the same and differ only in the functionality they provide – however, this is not the case. The obvious pointers are power consumption, POST time and latency. This can come down to the manufacturing process and prowess, so these are tested.

For Z490 we are running using Windows 10 64-bit with the 1909 update.

Power Consumption

Power consumption was tested on the system while in a single MSI GTX 1080 Gaming configuration with a wall meter connected to the power supply. This power supply has ~75% efficiency > 50W, and 90%+ efficiency at 250W, suitable for both idle and multi-GPU loading. This method of power reading allows us to compare the power management of the UEFI and the board to supply components with power under load, and includes typical PSU losses due to efficiency. These are the real-world values that consumers may expect from a typical system (minus the monitor) using this motherboard.

While this method for power measurement may not be ideal, and you feel these numbers are not representative due to the high wattage power supply being used (we use the same PSU to remain consistent over a series of reviews, and the fact that some boards on our testbed get tested with three or four high powered GPUs), the important point to take away is the relationship between the numbers. These boards are all under the same conditions, and thus the differences between them should be easy to spot.

Power: Long Idle (w/ GTX 1080)Power: OS Idle (w/ GTX 1080)Power: Prime95 Blend (w/ GTX 1080)

Long Idle and OS Idle power seem a lot higher than other Z490 boards. This will be down primarily to the PLX chip, but also the 10 gigabit Ethernet on board.

Non-UEFI POST Time

Different motherboards have different POST sequences before an operating system is initialized. A lot of this is dependent on the board itself, and POST boot time is determined by the controllers on board (and the sequence of how those extras are organized). As part of our testing, we look at the POST Boot Time using a stopwatch. This is the time from pressing the ON button on the computer to when Windows starts loading. (We discount Windows loading as it is highly variable given Windows specific features.)

Non UEFI POST Time

Despite not being a server board for Xeons, or having an IPMI, the Supermicro system has a similar POST time to those server boards. Part of this will be down to the PLX chip, but there is also consideration on CPU detection and training – a system can POST faster if it auto-assumes that the CPU and DRAM are the same as the last time it was turned on, whereas other motherboards will do a proper check every time.

DPC Latency

Deferred Procedure Call latency is a way in which Windows handles interrupt servicing. In order to wait for a processor to acknowledge the request, the system will queue all interrupt requests by priority. Critical interrupts will be handled as soon as possible, whereas lesser priority requests such as audio will be further down the line. If the audio device requires data, it will have to wait until the request is processed before the buffer is filled.

If the device drivers of higher priority components in a system are poorly implemented, this can cause delays in request scheduling and process time. This can lead to an empty audio buffer and characteristic audible pauses, pops and clicks. The DPC latency checker measures how much time is taken processing DPCs from driver invocation. The lower the value will result in better audio transfer at smaller buffer sizes. Results are measured in microseconds.

Deferred Procedure Call Latency

Normally anything under 250 microseconds is good, however it is clear that the other vendors are doing something that Supermicro is not.

For our motherboard reviews, we use our short form testing method. These tests usually focus on if a motherboard is using MultiCore Turbo (the feature used to have maximum turbo on at all times, giving a frequency advantage), or if there are slight gains to be had from tweaking the firmware. We put the memory settings at the CPU manufacturers suggested frequency, making it very easy to see which motherboards have MCT enabled by default.

For Z490 we are running using Windows 10 64-bit with the 1909 update.

Normally we test our motherboards with out of the box settings. This means that the performance will get boosted based on whatever default algorithm each motherboard vendor implements with regards turbo time and boost power. Intel actively encourages this – the numbers it puts in for turbo time and turbo power are recommendations, rather than specifications, and Intel wants motherboard vendors to engineer their products to the turbo and power that each vendor deems acceptable for their product. As a result, a lot of motherboards will implement an aggressive turbo algorithm.

For this generation, ASUS has done something different. ASUS’ enthusiast motherboards offer two different options on first boot: Intel recommendations, or ASUS recommendations. This means that there is a small performance delta between the two, especially for ASUS’ high-end motherboards. ASUS has put this into the product based on customer feedback and how motherboard vendors have slowly drifted over the last decade to well beyond what Intel recommends.

For our testing methodology, we try to leave as much as we can on default, because this is part of what makes a motherboard different to any other, and the motherboard vendor has to decide how aggressive it must be. Also, for non-enthusiasts who daren’t enter the BIOS, or understand even what turbo or a CPU or what memory channels are, they will just end up with the non-XMP default settings. It is unclear what such a person might select when presented with the ASUS default option.

On the side of Supermicro, as we’ve noted earlier in the review, whereas most vendors will adjust Intel’s recommendations for power limits and turbo, Supermicro appears to adhere to them strictly. This means some performance differential, more akin to ASUS’s option for Intel defaults.

Rendering – Blender 2.7b: 3D Creation Suite

A high profile rendering tool, Blender is open-source allowing for massive amounts of configurability, and is used by a number of high-profile animation studios worldwide. The organization recently released a Blender benchmark package, a couple of weeks after we had narrowed our Blender test for our new suite, however their test can take over an hour. For our results, we run one of the sub-tests in that suite through the command line – a standard ‘bmw27’ scene in CPU only mode, and measure the time to complete the render.

Blender 2.79b bmw27_cpu Benchmark

Rendering – POV-Ray 3.7.1: Ray Tracing

The Persistence of Vision Ray Tracer, or POV-Ray, is a freeware package for as the name suggests, ray tracing. It is a pure renderer, rather than modeling software, but the latest beta version contains a handy benchmark for stressing all processing threads on a platform. We have been using this test in motherboard reviews to test memory stability at various CPU speeds to good effect – if it passes the test, the IMC in the CPU is stable for a given CPU speed. As a CPU test, it runs for approximately 1-2 minutes on high-end platforms.

POV-Ray 3.7.1 Benchmark

Rendering – Crysis CPU Render

One of the most oft used memes in computer gaming is ‘Can It Run Crysis?’. The original 2007 game, built in the Crytek engine by Crytek, was heralded as a computationally complex title for the hardware at the time and several years after, suggesting that a user needed graphics hardware from the future in order to run it. Fast forward over a decade, and the game runs fairly easily on modern GPUs, but we can also apply the same concept to pure CPU rendering – can the CPU render Crysis? Since 64 core processors entered the market, one can dream. We built a benchmark to see whether the hardware can.

For this test, we’re running Crysis’ own GPU benchmark, but in CPU render mode. This is a 2000 frame test, which we run over a series of resolutions from 800×600 up to 1920×1080. For simplicity, we provide the 1080p test here.

Crysis CPU Render: (6) 1920x1080

Point Calculations – 3D Movement Algorithm Test: link

3DPM is a self-penned benchmark, taking basic 3D movement algorithms used in Brownian Motion simulations and testing them for speed. High floating point performance, MHz, and IPC win in the single thread version, whereas the multithread version has to handle the threads and loves more cores. For a brief explanation of the platform agnostic coding behind this benchmark, see my forum post here.

3D Particle Movement v2.1

NAMD ApoA1

One frequent request over the years has been for some form of molecular dynamics simulation. Molecular dynamics forms the basis of a lot of computational biology and chemistry when modeling specific molecules, enabling researchers to find low energy configurations or potential active binding sites, especially when looking at larger proteins. We’re using the NAMD software here, or Nanoscale Molecular Dynamics, often cited for its parallel efficiency. Unfortunately the version we’re using is limited to 64 threads on Windows, but we can still use it to analyze our processors. We’re simulating the ApoA1 protein for 10 minutes, and reporting back the ‘nanoseconds per day’ that our processor can simulate. Molecular dynamics is so complex that yes, you can spend a day simply calculating a nanosecond of molecular movement.

NAMD 2.31 Molecular Dynamics (ApoA1)

For Z490 we are running using Windows 10 64-bit with the 1909 update.

Grand Theft Auto V

The highly anticipated iteration of the Grand Theft Auto franchise hit the shelves on April 14th 2015, with both AMD and NVIDIA in tow to help optimize the title. GTA doesn’t provide graphical presets, but opens up the options to users and extends the boundaries by pushing even the hardest systems to the limit using Rockstar’s Advanced Game Engine under DirectX 11. Whether the user is flying high in the mountains with long draw distances or dealing with assorted trash in the city, when cranked up to maximum it creates stunning visuals but hard work for both the CPU and the GPU.

For our test we have scripted a version of the in-game benchmark. The in-game benchmark consists of five scenarios: four short panning shots with varying lighting and weather effects, and a fifth action sequence that lasts around 90 seconds. We use only the final part of the benchmark, which combines a flight scene in a jet followed by an inner city drive-by through several intersections followed by ramming a tanker that explodes, causing other cars to explode as well. This is a mix of distance rendering followed by a detailed near-rendering action sequence, and the title thankfully spits out frame time data.

GTX 1080: Grand Theft Auto V, Average FPSGTX 1080: Grand Theft Auto V, 95th Percentile

F1 2018

Aside from keeping up-to-date on the Formula One world, F1 2017 added HDR support, which F1 2018 has maintained; otherwise, we should see any newer versions of Codemasters’ EGO engine find its way into F1. Graphically demanding in its own right, F1 2018 keeps a useful racing-type graphics workload in our benchmarks.

Aside from keeping up-to-date on the Formula One world, F1 2017 added HDR support, which F1 2018 has maintained. We use the in-game benchmark, set to run on the Montreal track in the wet, driving as Lewis Hamilton from last place on the grid. Data is taken over a one-lap race.

GTX 1080: F1 2018, Average FPSGTX 1080: F1 2018, 95th Percentile

Strange Brigade (DX12)

Strange Brigade is based in 1903’s Egypt and follows a story which is very similar to that of the Mummy film franchise. This particular third-person shooter is developed by Rebellion Developments which is more widely known for games such as the Sniper Elite and Alien vs Predator series. The game follows the hunt for Seteki the Witch Queen who has arose once again and the only ‘troop’ who can ultimately stop her. Gameplay is cooperative centric with a wide variety of different levels and many puzzles which need solving by the British colonial Secret Service agents sent to put an end to her reign of barbaric and brutality.

The game supports both the DirectX 12 and Vulkan APIs and houses its own built-in benchmark which offers various options up for customization including textures, anti-aliasing, reflections, draw distance and even allows users to enable or disable motion blur, ambient occlusion and tessellation among others. AMD has boasted previously that Strange Brigade is part of its Vulkan API implementation offering scalability for AMD multi-graphics card configurations.

GTX 1080: Strange Brigade DX12, Average FPSGTX 1080: Strange Brigade DX12, 95th Percentile

Experience with the Supermicro C9Z490-PGW

Although overclocking to squeeze as much ‘free’ performance isn’t as beneficial as it was years ago due to both Intel and AMD doing the hard work for us via refinements to core architecture, it’s still an important aspect to consider. The reason behind this is these ‘refinements’ relate to Turbo and Boost clock speeds, which generally applies to one or two cores. This is great for single-threaded applications, which will benefit the most from higher clock speeds, but for multi-threaded applications and workloads, there’s more benefit to be had from overclocking all of the cores.

The main caveats to manually overclocking the processor as far as it can go is heat and power consumption, which can be negated with premium cooling solutions such as large AIO coolers. Intel’s Thermal Velocity Boost is also a variable to consider. With better cooling solutions, Intel’s Comet Lake processors can benefit with an extra 100 MHz on top of turbo, providing the processor is kept under 70°C. A hot processor will cause thermal throttling, which will harm system performance.

Our experience with the Supermicro C9Z490-PGW was generally pleasant. Still, the firmware itself, including the overclocking profiles, relies on Intel’s Default CPU specifications. Using the firmware, all of the overclocking related options can be found under the Overclocking tab within the Advanced section of the firmware, which can be accessed by pressing F7. There are many options for overclocking the CPU, including options for CPU Core ratio, BCLK frequency, and multiple voltage options.

Overclocking the memory is simple for users with kits that support X.M.P 2.0 profiles, which can done by simply enabling X.M.P. Users looking to manually overclock memory can do so with options for altering memory frequency and a variety of memory timings, as well as adjusting the DRAM voltage.

The best way to get the most out of a processor with the C9Z490-PGW is in the TDP configurations menu. By default, the board operates at Intel’s default settings, which means PL1 and PL2 limitations prevent the board from unleashing its capabilities. These options can also be found under the overclocking tab in a submenu called Config TDP Configurations. Altering higher than the default PL1 rating of 125 W will gain noticeable improvements in performance over stock settings, with an option to increase the time window to up to 128 seconds. 

Overall the Supermicro SuperO firmware is easy to navigate, and it is responsive. The only caveats we’ve found are highlighted below in our actual testing of the board’s overclocking capability. 

Overclocking Methodology

Our standard overclocking methodology is as follows. We select the automatic overclock options and test for stability with POV-Ray and Prime95 to simulate high-end workloads. These stability tests aim to catch any immediate causes for memory or CPU errors.

For manual overclocks, based on the information gathered from the previous testing, start off at a nominal voltage and CPU multiplier, and the multiplier is increased until the stability tests are failed. The CPU voltage is increased gradually until the stability tests are passed. The process repeated until the motherboard reduces the multiplier automatically (due to safety protocol) or the CPU temperature reaches a stupidly high level (105ºC+). Our testbed is not in a case, which should push overclocks higher with fresher (cooler) air.

Overclocking Results

The first thing that should be pointed out is that the Supermicro C9Z490-PGW is operating at Intel specification and doesn’t employ any profiles within the BIOS to rectify this. Whether or not this is an oversight on behalf of Supermicro, or they just wanted to leave it to a user’s own devices, this reflects heavily not only in our benchmark suite but in our overclock testing. This is similar to when we reviewed the ASUS ROG Maximus XII Hero WiFi, which offered users the option to choose between Intel specifications or ASUS’s own enhancements.

When we selected between the different profiles, including default, OC mode, and Autotuning, this made no real difference to performance in our POV-Ray benchmark. We saw a slight increase in both the latter profiles over the default settings, although not enough to warrant using them. We also noticed slightly higher than normal CPU VCore voltages at load compared to other boards we have tested, with very high load power consumption, which negates the point of adhering to Intel specifications.

Overclocking at Intel Default Specifications? Altering Power Limits = Performance

Even when we manually tested each frequency ranging from 4.7 to 5.2 GHz, we saw poor VDroop compensation when comparing the voltages set in the firmware to the CPU VCore under load. While we contemplated this due to sensor issues, our power consumption figures confirm it is drawing a lot of unnecessary power with no real benefit to performance. This shows that no matter what CPU VCore is set within the firmware, it makes no difference to the Intel default power settings, with no extra oomph from increased PL1 and PL2 limits that other boards use.

The only way we saw an extra increase in performance in line with other vendors was when we manually extended the power limits within the firmware. We tested this at 5.0 GHz with 1.275 V on the CPU VCore and saw a much better performance (*= manual power limit settings), especially compared to all the other Z490 models we’ve tested. This also shows that unless a user alters the power limit settings within the firmware, it will have no discernible impact on performance, no matter the CPU VCore and CPU Frequency settings that are inputted.

The other caveat is the board’s VDroop control is some of the poorest we’ve seen on Z490 so far, which caused our chip to throttle badly at 5.2 GHz and give very high power consumption figures compared to other Z490 models. The TL:DR is unless you alter power limits, performance is handicapped. Unlike other boards, simply doing overclocking adjustments doesn’t automatically change those values.

A lot more focus has been put onto power delivery specifications and capabilities, not just by manufacturers, but as a result of users demands. In addition to the extra power benefits from things like overclocking, more efficient designs in power deliveries and cooling solutions aim to bring temperatures down. Although this isn’t something most users ever need to worry about, certain enthusiasts are bringing more focus onto each boards power delivery. The more premium models tend to include bigger and higher-grade power deliveries, with bigger and more intricate heatsink designs, with some even providing water blocks on ranges such as the ASUS ROG Maximus Formula series.

The 8+2 power delivery on the Supermicro C9Z490-PGW

Testing Methodology

Our method of testing out if the power delivery and its heatsink are effective at dissipating heat, is by running an intensely heavy CPU workload for a prolonged method of time. We apply an overclock which is deemed safe and at the maximum that the silicon on our testbed processor allows. We then run the Prime95 with AVX2 enabled under a torture test for an hour at the maximum stable overclock we can which puts insane pressure on the processor. We collect our data via three different methods which include the following:

  • Taking a thermal image from a birds-eye view after an hour with a Flir Pro thermal imaging camera
  • Securing two probes on to the rear of the PCB, right underneath CPU VCore section of the power delivery for better parity in case a probe reports a faulty reading
  • Taking a reading of the VRM temperature from the sensor reading within the HWInfo monitoring application

The reason for using three different methods is that some sensors can read inaccurate temperatures, which can give very erratic results for users looking to gauge whether an overclock is too much pressure for the power delivery handle. With using a probe on the rear, it can also show the efficiency of the power stages and heatsinks as a wide margin between the probe and sensor temperature can show that the heatsink is dissipating heat and that the design is working, or that the internal sensor is massively wrong. To ensure our probe was accurate before testing, I binned 10 and selected the most accurate (within 1c of the actual temperature) for better parity in our testing.

To recreate a real-world testing scenario, the system is built into a conventional desktop chassis which is widely available. This is to show and alleviate issues when testing on open testbeds which we have done previously, which allows natural airflow to flow over the power delivery heatsinks. It provides a better comparison for the end-user and allows us to mitigate issues where heatsinks have been designed with airflow in mind, and those that have not. The idea of a heatsink is to allow effective dissipation of heat and not act as an insulator, with much more focus from consumers over the last couple of years on power delivery componentry and performance than in previous years.

For thermal image, we use a Flir One camera as it gives a good indication of where the heat is generated around the socket area, as some designs use different configurations and an evenly spread power delivery with good components will usually generate less heat. Manufacturers who use inefficient heatsinks and cheap out on power delivery components should run hotter than those who have invested. Of course, a $700 flagship motherboard is likely to outperform a cheaper $100 model under the same testing conditions, but it is still worth testing to see which vendors are doing things correctly. 

Thermal Analysis Results


We measured 86.4°C on the hottest part of the PCB around the power delivery

The Supermicro C9Z490-PGW is using a 10-phase power delivery which is controlled by an Infineon XDPE12284C PWM controller operating at 8+2. The power delivery is using eight Infineon TDA21490 90 A power stages for the CPU, with two Infineon TDA21535 70 A power stages for the SoC  It is cooled by a pair of aluminum heatsinks, with more focus on the surface area than relying on mass. This type of design requires a chassis with good passive airflow.

Focusing on the thermal performance of the Supermicro C9Z490-PGW’s power delivery, and it’s certainly not the coolest we have encountered. Despite including a large passively cooled dual heatsink, it struggles to cope with heat under full-load when overclocked. We measured readings of 78°C and 79°C respectively from our pair of K-type thermocouples, while this board doesn’t benefit from an integrated thermal sensor.

Using our thermal imaging camera, we took a reading of 86.4°C from the hottest part of the PCB around the CPU socket area. The cooling properties of the heatsink on the C9Z490-PGW isn’t as efficient as other ATX sized Z490 models we have tested, but the temperatures are still well within the official specifications. 

Albeit more known for its workstation models, Supermicro isn’t as famous for its gaming-focused models. Last year we reviewed the Supermicro C9Z390-PGW, which was competitive with other Z390 models. This year when Intel unveiled its 10th generation Comet Lake processors, Supermicro unveiled its Z490 offering, with a similar feature set and close pricing to its predecessor. Dropping its ‘Play Harder’ moniker for Z490, Supermicro instead markets this board for ‘Ultimate HEDT Performance,’ and although socket LGA1200 isn’t technically HEDT, it does support the Core i9-10900K, which has 10 cores and 20 threads. 

Diving right into the finer specifics of the Supermicro C9Z490-PGW, its most marketable feature is the inclusion of a Broadcom PEX8747 PLX chip. This allows the PCI lanes from the CPU to be muxed, which adds extra functionality regarding PCIe expansion. There are four full-length PCIe 3.0 slots that can be used at either x16/-/x16/- or x8/x8/x8/x8. Back in the day, this was a solid solution for users with 4-way NVIDIA SLI setups. Still, in today’s current market, both NVIDIA and AMD are moving away from multi-graphics support and instead opting for a solid single graphics solution. This does open the door to using it for professional use without going down the W480 route, with plenty of point-to-point bandwidth between add-in-cards for extra storage and RAID controllers, as well as additional networking controllers. 


Broadcom PEX8747 PLX chip: Muxing it up

The other elements of the C9Z490-PGW position it under the premium category, with a solid networking array led by an Aquantia AQC107 10 GbE controller, with an assisting Intel I219-V Gigabit PHY. Wireless connectivity is taken care of by an Intel AX201 Wi-Fi 6 interface, which also adds BT 5.1 support. The board’s integrated audio is a premium yet standard for Z490 with a Realtek ALC1220 HD audio codec, which powers five 3.5 mm audio jacks and S/PDIF optical out on the rear a front panel audio header located underneath the codec on the PCB. Other notable inclusions on the rear panel include a USB 3.2 G2x2 Type-C port, three USB 3.2 G2 Type-A, one USB 3.2 G2 Type-C, two USB 3.2 G1 Type-A ports, as well as a DisplayPort 1.4 and HDMI 2.0a video output pairing. 

Looking at performance, and straight away, it is apparent that the Supermicro is adhering to Intel specifications and not relying on its own ‘sauce’ within the firmware. This does hinder performance somewhat when Intel’s PL1 and PL2 limits are taken into consideration compared to other boards, but is perhaps closer to Intel’s factory vision (which may be different to Intel’s PR vision). We saw slower than usual POST times for a Z490 model in our system tests – this is normal for a Supermicro board, but it means that Supermicro isn’t taking the shortcuts in firmware initiation that the other vendors do. Power consumption, on the other hand, is a tad high both at default and when overclocking, due to the PLX chip – acceptable but something to note. It also includes support for up to DDR4-4000 memory which is lower than the majority of Z490 models on the market, but Supermicro offers functionality rather than memory going way beyond manufacturer support. 

When manually overclocking on the C9Z490-PGW, making power limit adjustments proved vital as when left at default settings, they will limit performance even at 5.0 GHz all-core. The firmware does allow for manual adjustment of these, and as seen in our overclocking testing, it is the limiting factor to unlocking the maximum potential of this board. When setting the CPU VCore within the BIOS at full load, VDroop is overcompensating heavily when it doesn’t need to be, increasing both temperatures and power consumption. It uses a solid 8+2 phase power delivery, which is more than capable, but this is likely down to the firmware. Even when setting the LLC levels, VDroop was still a little awry, and it could be massively improved upon.

 

Final Words

As it stands, the Supermicro C9Z490-PGW offers more in the way of functionality than other Z490 models, largely due to the inclusion of a PLX chip to mux the four full-length PCIe slots. It’s also the cheapest Z490 model to include 10 GbE networking with an MSRP of $395, with the next model up, the ASUS ROG Maximus XII Formula costing $475. From this perspective, the Supermicro C9Z490-PGW, with everything it has to offer, including PLX, 10 gigabit Ethernet, Wi-Fi 6, and solid 10-phase power delivery, it’s a good overall package. Users familiar with adjusting Comet Lake’s PL1 and PL2 settings (not hard to figure out) can get themselves a good board for an equally good price. Those not too bothered about Wi-Fi can opt for the Supermicro C9Z490-PG, which is the same board minus the Intel AX201 CNVi, which is currently $343 at Newegg.