Discussion in 'The Veterans' Lounge' started by daknife56, Apr 18, 2015.
I have to agree with majority except when i zone to pok i change zone within a few seconds pok about 20 seconds and thats with a very average pc i7 1st gen 3.06ghz 6gb of RAM. My internet connection is only around 80mbps. From character select to enter the zone is about same time as for zoning to pok.
Time to eat some humble pie...
I've just run some zoning tests, running the following path: Character Select - Lobby - PoK - Lobby - PoK - PoT - PoK - Lobby
Ran the test on identical installations of EQ, one on a SSD, one on a rotating HDD (copied the EQ folder from one to the other). Memory mode set to Balanced.
Did a full defrag on all drives before starting the first test. Rebooted the PC before each test. During each test, the only application running was the current install of EQ being tested. (plus usual background services, antivirus, etc)
Took the timings from the character log files (precise only to one second), from the "LOADING PLEASE WAIT" message to the "You have entered... " message. Also took stopwatch timings from the moment the client stopped responding following clicking the lobby door/PoK stone (i.e. could no longer move), until the next zone was displayed on the screen.
Using the log file timings:
Running from the HDD installation, the journey took 69 seconds (total zoning times)
On the SSD, the journey took 63 seconds (total zoning times).
So, on my current system (i5-4430, 8GB, WD20PURX-64P6ZY0 HDD), the difference is only about 10%.
Having said that, on my previous system (where a single zoning usually took well over a minute), the performance boost was from something in the order of 1m 35sec down to 20s (hence the original figure of 5x faster). Unfortunately I can't remember the specs of the previous system, but safe to say that every component was inferior to the current rig, and, presumably, the transfer speed from the HDD was sufficiently bad that the SSD made such a significant difference.
Also worth mentioning that the observed zoning times (using a stopwatch) were usually around double that reported in the log files (total 164s for the HDD test, compared to 69s reported in the log files), so I'm guessing there is some latency before EQ writes the "LOADING..." message to the file, and/or the "You have entered..." message is logged before the display is updated. Using these times, the total figures are 164s for HDD, 151s for SDD (approx 8% difference).
So, if you've got an older system, you may see more of a benefit with an SSD, but for any recent system, I have to concur with Xanadas, it appears to make not much of a difference any more.
Anyone remember those few months where (I want to say during Velious era but not sure) texture caching was used to speed up zoning? The first time you zoned, it took a bit longer than normal because it was building the cache files, but then after that -- subsequent zonings took seconds (and this was on hardware of that time that was not so hot by today's standards).
I think there were issues created when zone files were changed after a patch (so then you had to look for and delete these cache files so that they could be rebuilt again) and probably in that light, caused more trouble than it was worth (so it was discontinued) -- but maaaaan, you zoned like lightning!
You could replace load times with just about anything dealing with EQ from a technical standpoint and be correct. It's people (primarily) and content (secondarily but highly important) that keep it going.
Really if nothing is happening on your connection at home and people are not downloading stuff you're not watching netflix. EQ plays the same on a 80mbps as it does on a 10 mbps link assuming you have the same latency.
Now in some other games where latency is a bigger issue like a FPS or league of legends if the 10 mbps link has a latency of 40ms to the game server and the 80 mbps has a latency of 80ms to the game server the 10mbps link is much much better.
Usually you should be worried about which ISP performs better rather than who has bigger bandwidth packages, a good example is if you have the choice between a 6mbps dsl link or a 12 mbps fixed wireless (4g wimax or lte) the DSL is likely going to have a much better latency 9 times out of 10 and it would be better for gaming but if all you do is download things and watch netflix the wireless would be better.
TLDR; 80mbps internet connection doesn't mean anything in most video games (landmark , second life and a few others benefit from being able to download lots of junk)
There's lots of other variables that can effect load times which can make the SSD appear better also like if you're boxing and all of the accounts have logs on you start to see the SSD pull ahead of the hdd.
Question in your testing was the OS installed on the same HDD and on the same SSD as the EQ instance? A HDD that only has EQ on it and no os or anything else on it isnt going to be too far behind the SSD but if you're running the OS on the same hard drive as you play video games on which is the case for many people the SSD is going to start to give more of a performance increase over the HDD.
Also 10-20% isn't bad
Since the April 22nd patch I think zoning times (especially to our GGH) have improved a lot. Nothing changed on my end but zoning times seem cut in half.
From a non technical point of view with no numbers to back it up. I have always seen better EQ load times with a new Video card upgrade and not the other parts of the system. I think the double load times you are seeing is a result of maybe that. If you are in a group with someone and port to the same location you can always see them port in, then vanish and show back up. At that point they seem to be able to move and talk. Though when you first see them pop in you can send a tell and not get the Queued message even though on their end they have not seen the message yet or unable to respond.
Separate names with a comma.