HP Microserver 1: AMD Athlon II Neo N36L 1GB RAM 160GB (P/N: 612275-421)

I finally got round to benchmarking some Graphics cards on the HP Microserver. Although I have not played around with much benchmarking before I thought I would start here as it seems to be a hot topic at present.

My first thoughts were “What do people really want out of the HP Microserver?” – Originally I wanted it as a replacement to my Acer Aspire Revo 3600 and run it as my dedicated HTPC. However this developed through time to become a replacement for my NETGEAR MS2000 Stora NAS while streaming the shares and content to my Acer Aspire Revo 3600 running XBMC.

I was previously using the Negear Stora, but was not satisfied that if the drive failed that I could easily get the data from one of the mirrored disks due to its XFS file system. There are some guides to doing this and I played around with accessing the XFS HDD from a Linux VM but this got quite messy and I had to refer to the good old rule “K.I.S.S.” therefore decided to purchase 2 x 2TB hard drives for the Microserver, RAID1 them and then dump all my media onto here and share using the Dreamspark version of Windows Server 2008 with shared folders. A nice method, however this didn’t allow me to use Streaming Media to my DNLA enabled devices, TV, iPhone, WDTVLive & other Windows 7 Computers.

With the release of Windows Home Server 2011 and dedicated DNLA support I decided I would install this as a VM on top of Windows Server 2008 R2 – 100% NOT SUPPORTED and definately frowned upon by many. But it is all in the name of testing and experimenting with the system. The reason I did this was that in order that I was still able to play around with Hyper-V and Windows Server 2011.

HP Microserver Number 2: AMD Athlon II Neo N36L 1GB RAM 250GB (P/N: 633724)

I recently purchased my 2nd Microserver to play around with Citrix Xen Server and would allow my current Windows Server/Home Server 2011 to stay in one piece. Xen works fine on this box, however I have not put it to the test and only had a limited number of VM’s on the system so will look at the development of this later on.

Following on from my original article I decided to see how well the performance was of the Radeon HD5450 Graphics Card. There has been a lot of talk about whether this card will do a good job as it seemed to be the fastest low profile card (at time of this article) that could be found. To try and improve performance I decided to run with Windows 7 x64 as the base OS, this offered fully supported drivers for all the graphics cards.

Windows 7 was installed on an external HDD and attached using the eSATA connection. A good solution if you don’t want to mess around with the current drives and OS already installed.

I ran through a number of tests on the following cards:

1. Original Radeon 4200 (This is the onboard graphics card) – VGA with no Audio
2. PNY Geforce 8400GS (Low Profile PCIe) with Heatsink – HDMI with Audio
3. Sapphire HD5450 (Low Profile PCIe) with Heatsink – HDMI with Audio

I used the following tools for benchmarking: (There maybe better tools available however as I have only played around with this brief this is what I went with)

1. Passmark Performance Test 7.0 (Trial Version)
2. 3D Mark Vantage (Free Basic Version) The latest version of 3D Mark 11 only works with Directx 11 cards.
3. PC Mark 7 (Free Basic Version)
4. CPUID (Free)

Passmark Performance Test Scores

1. Original Radeon 4200 – Score: 440.9 

   no images were found

2. PNY Geforce 8400GS – Score: 455.6 

   no images were found

3. Sapphire HD5450 – Score: 468.7 

   no images were found

3D Mark Vantage

1. Original Radeon 4200 – Score: TBA
2. PNY Geforce 8400GS – Score: P435

   no images were found

3. Sapphire HD5450 – Score: P1091

   no images were found

PC Mark 7

1. Original Radeon 4200 – Score: 974 
   
2. PNY Geforce 8400GS – Score: 994 
   
3. Sapphire HD5450 – Score:1034
   

CPUID Temperature Tests

1. Original Radeon 4200 N/A
2. PNY Geforce 8400GS 
   
3. Sapphire HD5450 
   

Power Tests:

Another question I always asked myself was whether the Microserver could be left on all day without costing too much and eating up too much power? The results from a electric meter gave a good indication to this and looks similar to leaving a couple of lights on.

Wattage when Idle (GF8400GS Card, 1 x eSATA & 2 x SATA HDD)

no images were found

Wattage when Underload (GF8400GS Card, 1 x eSATA & 2 x SATA HDD)

no images were found