Upgrading my PC for SWTOR

[Max_Killjoy]

The first thing I did with my current rig, years ago, was go from 4GB to 8GB.

[skarvika]

Honestly the parts you use aren't going to matter too much. This game's engine is bantha poodoo, and you will lag.

Your current setup is good enough though so I wouldn't worry about adding new parts yet.

[TX_Angel]

Gonna agree with TX here except to say 16GB is my floor and has been since Windows 7 64 bit was stable.

 

Well I tend to agree personally, my main machine has 32GB and has for several years. There have been times I've used most of it as well.

 

But I also try to take into account other people don't use their computer the way I do. If you're talking a $400 entry level machine, going to 16GB might not be the first place to put money into it.

 

Example:

 

ASUS M32CD Desktop - $450

http://amzn.to/1KZMQ0q

 

That computer has an Intel Quad Core i5-6400 Skylake CPU, 8GB of DDR4 RAM, and a 1 TB hard drive.

 

If you came to me and said "what is the ONE upgrade that will make the BIGGEST and most NOTICBLE difference", it would not be to upgrade it to 16GB of RAM.

 

It would be to put a SSD in there, cloning the existing 1TB hard drive to it and making the 1TB drive a data storage drive.

 

It makes a monster difference in terms of overall system snappiness and performance.

 

Of course, that machine also needs a video card for gaming, which is a separate issue. A GTX 750 TI actually would work well in that machine, since it doesn't need a PCI-E 6 pin power cable, however a GTX 950 would also work well enough, since the machine has a pair of molex connectors and a EVGA GTX 950 comes with a dual molex to PCI-E 6 pin adapter cable.

 

EVGA GTX 750 TI - $110

http://amzn.to/20JGpkf

 

EVGA GTX 950 - $150

http://amzn.to/1Qpw4dG

 

DDR4 8GB - $37

http://amzn.to/1KZNc7x

 

DDR4 still costs a premium over DDR3, which is a shame, since it doesn't really matter for most home users. 8GB DDR3 is $20, right now 8GB DDR4 is $37.

 

I would suggest a 240GB SSD long before a 8GB to 16GB RAM upgrade:

 

Crucial BX200 240GB SSD - $65

http://amzn.to/1oQabYL

 

There are slightly cheaper options (by like $5 or so), but I'd pay the extra $5 to get a Crucial drive. I own multiple of the above drive and they work perfectly in every machine I've put them into. That is well worth $5 to me.

 

SWTOR is a 32 bit application. It will NEVER address more than 2^32 bits (4GB) of memory. If your system is running only 4 GB of memory, you have to account for EVERYTHING that is running then tack on whatever else you want to run. This is assuming you are even running a 64 bit OS (it could happen). I can't say that I've ever seen it running more than 1.5-2GB at any given time that I've looked though. Granted I don't keep a monitor on it all the time.

 

Actually, 32-bit Windows programs generally can't access over 2GB, due to the way 32-bit Windows handles user space. It set aside (a long time ago) 2GB for user programs and 2GB for the system.

 

SWTOR runs 2 processes to get around this. The first often bumps right up into the 2GB limit, with the second one often using 500MB to 1GB of RAM, give or take.

 

Still, even at less than 3GB used, there is data that Windows caches from the hard drive, both for the game and for everything else. Then there are the 15 things most users have running in their task bar. And many people like to play in fullscreen (windowed) mode and alt tab over to a web browser, which takes more RAM these days than you'd think.

 

All that being said, I consider 8GB to be the minimum, but 16GB is my suggested amount for many people who either will play games, or want to have 5+ things left open at once. And if the price of the computer goes north of $600 or so, I see no reason to not have 16GB in there, the percentage of the total system cost vs. the price of the RAM just makes less seem like stepping over dollars to save pennies to me.

[MeNaCe-NZ]

Just want to point out that on my 4690k I was able to overclock from 3.5GHz to 4.4GHz on the Hyper 212 EVO. That's a 25% clock increase.

 

Yeah I got to 4.3 on mine - the 6600Ks look like they have a better overall average overclock too.

[Draqsko]

Sure there is! Get a better computer!

 

Keep in mind I was running those tests on a 9 year old machine. SWTOR was released in 2011, but it has been updated many times and the game that existed at launch isn't the game we're playing today.

 

I see people playing on machines that are older than the launch of the game (someone else in the other thread is running a AMD Phenom II X4 925, that CPU came out 1Q 2009, it is 7 years old. Expecting it to perform well in 2016 in an updated MMO is not very reasonable, IMHO. He appears to be happy with it, so more power to him, however it is what is keeping him from running smoother and faster. He doesn't PvP however so perhaps it doesn't matter to him.

 

http://www.cpu-world.com/CPUs/K10/AMD-Phenom%20II%20X4%20925%20-%20HDX925WFK4DGI%20(HDX925WFGIBOX).html

 

 

 

I'm in Dallas, TX sadly, far away from both sets of servers. My ping is pretty solid at around 50ms, give or take 5. I have fiber to the home (AT&T Gigafiber) and office (Verizon FIOS) and the hop time is low, but the distance is not and I can't change the physical distance between Texas and the servers. Electricity only moves so fast.

 

I meant besides upgrading your computer , but yes the game requirements have gone up over the years. When I started playing SWTOR, I was playing on an Athlon 64 X2 6400+ oc'd to 3.5 GHz, and 2x ATI Radeon HD 2600 XT crossfired. I could get 30 FPS on medium settings in a warzone although open world would just cause it to tank once you had over 20 people around you fighting. A year ago before I built the current rig I have, I was only getting single digit FPS in warzones with a Radeon HD 6770 and the same CPU (although the IMC was failing at that point, it was only detectable with a stress test). Largely it's due to the increase in particle effects in the game which of course puts a heavy load on your CPU.

 

That's actually a pretty good ping for your distance, I get roughly that with Tera given their servers are in Chicago but I still occasionally get packet loss at times (especially prime time since internet usage in my area goes up dramatically once everyone gets home). When that happens you can really tell in Tera because you'll desync with your attacks and pass through your target when you shouldn't, or your inputs will feel less responsive. It's frustrating as hell but it's largely beyond yours or the developer's control since it's depending on the internet infrastructure between you and the server. In my case I can safely blame Level3 since their equipment is where the packet loss or packet delay occurs nearly all the time. Unless you run something like pingplotter though, it'll be very hard to actually see where it's happening.

[TX_Angel]

That's actually a pretty good ping for your distance, I get roughly that with Tera given their servers are in Chicago but I still occasionally get packet loss at times (especially prime time since internet usage in my area goes up dramatically once everyone gets home). When that happens you can really tell in Tera because you'll desync with your attacks and pass through your target when you shouldn't, or your inputs will feel less responsive. It's frustrating as hell but it's largely beyond yours or the developer's control since it's depending on the internet infrastructure between you and the server. In my case I can safely blame Level3 since their equipment is where the packet loss or packet delay occurs nearly all the time. Unless you run something like pingplotter though, it'll be very hard to actually see where it's happening.

 

When I ping the west coast servers:

 

Pinging 159.153.68.252 with 32 bytes of data:

Reply from 159.153.68.252: bytes=32 time=48ms TTL=242

Reply from 159.153.68.252: bytes=32 time=47ms TTL=242

Reply from 159.153.68.252: bytes=32 time=47ms TTL=242

Reply from 159.153.68.252: bytes=32 time=47ms TTL=242

 

Ping statistics for 159.153.68.252:

Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),

Approximate round trip times in milli-seconds:

Minimum = 47ms, Maximum = 48ms, Average = 47ms

 

I don't know that I can do much better, being that I'm nearly 2,000 miles from the server (who knows the actual distance the wires take of course).

 

I ran a TraceRT and I'm exactly 12 hops to the server from my router:

 

Tracing route to 159.153.68.252 over a maximum of 30 hops (some IPs edited out)

 

1 <1 ms <1 ms <1 ms dsldevice.attlocal.net [192.]

2 6 ms 7 ms 106 ms 104-lightspeed.rcsntx.sbcglobal.net [104.

3 3 ms 2 ms 2 ms 71.

4 6 ms 5 ms 5 ms 70.143.

5 7 ms 5 ms 5 ms 70.143.

6 9 ms 7 ms 7 ms 12.83.

7 9 ms 7 ms 7 ms gar26.dlstx.ip.att.net [12.123.16.85]

8 * * * Request timed out.

9 * * * Request timed out.

10 * * * Request timed out.

11 47 ms 47 ms 47 ms 4.28.172.102

12 48 ms 47 ms 47 ms 159.153.68.252

 

I don't think I can improve that much...

[Draqsko]

When I ping the west coast servers:

 

Pinging 159.153.68.252 with 32 bytes of data:

Reply from 159.153.68.252: bytes=32 time=48ms TTL=242

Reply from 159.153.68.252: bytes=32 time=47ms TTL=242

Reply from 159.153.68.252: bytes=32 time=47ms TTL=242

Reply from 159.153.68.252: bytes=32 time=47ms TTL=242

 

Ping statistics for 159.153.68.252:

Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),

Approximate round trip times in milli-seconds:

Minimum = 47ms, Maximum = 48ms, Average = 47ms

 

I don't know that I can do much better, being that I'm nearly 2,000 miles from the server (who knows the actual distance the wires take of course).

 

I ran a TraceRT and I'm exactly 12 hops to the server from my router:

 

Tracing route to 159.153.68.252 over a maximum of 30 hops (some IPs edited out)

 

1 <1 ms <1 ms <1 ms dsldevice.attlocal.net [192.]

2 6 ms 7 ms 106 ms 104-lightspeed.rcsntx.sbcglobal.net [104.

3 3 ms 2 ms 2 ms 71.

4 6 ms 5 ms 5 ms 70.143.

5 7 ms 5 ms 5 ms 70.143.

6 9 ms 7 ms 7 ms 12.83.

7 9 ms 7 ms 7 ms gar26.dlstx.ip.att.net [12.123.16.85]

8 * * * Request timed out.

9 * * * Request timed out.

10 * * * Request timed out.

11 47 ms 47 ms 47 ms 4.28.172.102

12 48 ms 47 ms 47 ms 159.153.68.252

 

I don't think I can improve that much...

 

Well that's why I said Pingplotter, since it's far better at tracing than tracert and pathping.

 

https://www.pingplotter.com/

 

You can even use it while you are playing, although it will cause some latency when it runs while you are playing.

 

PS. Hops 8, 9, and 10 could be an issue since they are timing out. Generally that happens when the routers are overloaded and ping requests are dropped in favor of actual data packets.

Edited February 20, 2016 by Draqsko

[TX_Angel]

Well that's why I said Pingplotter, since it's far better at tracing than tracert and pathping.

 

https://www.pingplotter.com/

 

You can even use it while you are playing, although it will cause some latency when it runs while you are playing.

 

PS. Hops 8, 9, and 10 could be an issue since they are timing out. Generally that happens when the routers are overloaded and ping requests are dropped in favor of actual data packets.

 

I just downloaded and ran that program, thanks for the suggestion.

 

Hop 8, 9, and 10 are Level3, they won't generally return pings for whatever reason. 4.69.152.89 is the IP address to that router which is in San Jose, CA.

 

What is interesting is that almost all the hops are out in CA. According to both TraceRT and PingPlotter, the jump from the Dallas, TX area to Los Angeles, CA happens at hop 3, so it appears that my local router connects directly to the main backbone between Dallas and Los Angeles, there is no hop between them.

 

It ALSO appears that all of the delay is at Level3, since the jump out to Los Angeles is actually really quick, it shows just 5ms for that, which seems REALLY low given the nearly 2k mile distance.

 

The first jump over 10ms shows up at hop 9 at Level3 out in San Jose, CA.

 

---

 

Now that being said, as I watch this, it is interesting... My local IP connection doesn't appear very clean. It is in the 2-4ms range "most of the time", but it does jump up higher from time to time.

 

It appears that the issue is "jitter", which I've heard of before but never looked at before.

[TX_Angel]

Update:

 

When I run this site:

 

http://www.pingtest.net/

 

It says my local ping is 7ms and my jitter is 1ms, connecting to a local server in Dallas.

http://www.pingtest.net/result/140273616.png

 

If I pick a server out in Los Angeles I get a ping of 35ms and jitter of 1ms:

http://www.pingtest.net/result/140273653.png

 

If I pick a server out of Orlando, FL, I get a ping of 39ms and jitter of 1ms:

http://www.pingtest.net/result/140273688.png

 

Just for fun, I tried Brisbane, Australia, where I get a ping of 203ms and jitter of 1ms:

http://www.pingtest.net/result/140273723.png

 

Not bad considering a distance of 8,250 miles.