Sir Cabirus' Scene load time measurement thread

Hello, I did some time-checks how long it takes to reach the overland map from a town or vice versa. Although I have published some of this measurement results already here, I have decided to create a own thread. If this hurts a forum rule please simply merge this thread.

You find the results of my measurements here, for convenience reasons here is a sceenshot too



I tried to disable all other network traffic (like Dropbox synchronization, driver and system updates and so on) during the measurements. I had some unexplainably anomalies which doesn't fit into the other results. But a result is a result and I decided to use these anomalies nevertheless. As you may know I'm a German living in Germany, therefore these are the results of an 'overseas measurement'

All times were measured with my iPhone 6 build-in hand-held stopwatch. I have not enough spare time to make all these measurements during one day. In fact I needed 4 days which may have an influence on the results because of potentially different server workloads. So my results are only indications.

Each measuring point has four manifestations

1. SotA installed on a SSD, measurement in single player online (SSD SPO)
2. SotA installed on a SSD, measurement in multi player online (SSD MO)
3. SotA installed on a HDD, measurement in single player online (HDD SPO)
4. SotA installed on a HDD, measurement in multi player online (HDD MO)

Each measurement was executed five times. The value of each of these measurements and the resulting average time are shown in the table. Anomaly results are marked in red.

Conclusion:

• It doesn't seems to matter whether the game is installed on a SSD or a HDD
• There is no significant difference between SPO and MO mode

My equipment:
CPU: QuadCore Intel Core i7-2600K, 3500 MHz (35 x 100)
Motherboard: Asus P8P67 (3 PCI, 2 PCI-E x1, 2 PCI-E x16, 4 DDR3 DIMM, Audio, Gigabit LAN, IEEE-1394)
RAM: 16360 MB (DDR3-1333 DDR3 SDRAM)
Videocard: NVIDIA GeForce GTX 570 (1280 MB)
SSD: Samsung SSD 840 EVO 500G (465 GB, SATA-III) <- here is SotA installed on SSD
SSD: M4-CT256M4SSD2 (256 GB, SATA-III)
HDD: WDC WD10EZEX-00BN5A0 (1000 GB, 7200 RPM, SATA-III) <- here is SotA installed on HDD
HDD: Hitachi HDS722020ALA330 (2000 GB, 7200 RPM, SATA-II)
OS: Microsoft Windows 7 Professional, Service Pack 1

 

Nice work!

 

The one thing I am seeing that might make a huge difference is the choice of which SSD SotA was running from.

I would see what the difference is with it installed on the Crucial SSD if possible, as you are effectively comparing apples to oranges by gimping the results on a IDE SSD.

Other than that one CC, I appreciate what you've done here!

 

It is interesting that most of your anomalies are in the first measurement of any set (E.g. Kingsport - Ardoris at 115 seconds is your first measurement).

I want to say that there is a correlation there. When you did a single set of loads (e.g. Kingsport - Ardoris SSD SPO) did you do them sequentially with the game still open? I can understand closing / quitting the game between the different sets e.g. SSD SPO and SSD MO. The fact that the times get better after the first attempt makes me want to guess that the data is held in ram from the previous attempt.

Good work though.

 

Looks kinda like the guards gave Sir Cabirus a hard time when he was trying to enter a city (maybe has to do with your reputation of causing the guards problems.....hmmmmmmm, maybeeeeee.

Then after going to a tavern, having a few ales, the guards were happy to show you the city gates?

Nice work SC.

 

Thank you for your hint, Lord Tachys al`Fahn. The IDE declaration is an error, I simply did copy this information from my AIDA64 Extreme Edition tool but IDE is a wrong specification



In fact my motherboard offers no IDE connections My Samsung 840 Evo is a SATA III ((6Gb/s) drive. You can look at the specifications here, if you like.

Nevertheless I made some comparison measurements between both SSDs



As expected there a no differences between these drives. But the SPO mode is around 5 seconds faster now.

Thank you again for your hint, I have corrected my mistake in my equipment specs.

 

This is a very valuable hint Jatvardur. Thank you for your statement

I think your assumption is true, this must be a kind of caching effect. I made some further measurements which seems to endorse this theory.

 

All righty then! I would have thought that there might be more of a difference between even a 7200 rpm HDD and an SSD, even on SATA-III, but there you are!

Good stuff!

 

A SSD drive is much faster as a HHD drive of course, but this is not the point. The game reads a mixed data stream to build up a scene. One part is from the local drive, the other part is from the net. Both parts must be read completely before a scene can be shown. If the net part is slower than the data stream from the drive, the drive speed doesn't matter.

 

And likely enough there is! Just that you see non of it because there's a bottleneck somewhere, hence all you see is the limiting factor. Others that do not suffer from that one chokepoint (whatever it may be) would likely report different figures.

Edit: Hmm, I should have posted faster, didn't see Sir Cabirus had already replied, sorry.