User panel stuff on forum
  55 posts on 2 pages  First page12Last page
Help
2010-12-24, 01:58
Member
284 posts

Registered:
Oct 2006
Rikoll wrote:
The fps drop from dying, is because the scoreboard is drawn ye?

Well that can probably be why in a sense, but does it sound reasonable that if I'm running at about 2800 fps at dm3 rl (localhost, alone) and bore myself to the ground, when the scoreboard is shown, my fps dips to something like 1300 fps? That's a 1500 fps drop because of scoreboard! Doesn't sound reasonable to me :E
2010-12-24, 03:22
Member
284 posts

Registered:
Oct 2006
Ok,

So this cfg couldn't do stable 1000 fps, but I have to remind you that I had g_weapons installed AND I didn't have threaded optimization on. It seems to hold 1000fps now, but not 1200 fps.

I tried to strip it as little as possible in vague areas.

////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////
/// ///
/// ///
/// E Z Q U A K E C O N F I G U R A T I O N ///
/// ///
/// ///
////////////////////////////////////////////////////////////////////////////////////////////////////
////////////////////////////////////////////////////////////////////////////////////////////////////



////////////////////////////////////////////////////////////////////////////////////////////////////
// //
// C O M M A N D L I N E //
// //
////////////////////////////////////////////////////////////////////////////////////////////////////


// ezquake-gl.exe

////////////////////////////////////////////////////////////////////////////////////////////////////
// //
// V A R I A B L E S //
// //
////////////////////////////////////////////////////////////////////////////////////////////////////


//All variables (even those with default values) are listed below.
//You can use "cfg_save_unchanged 0" to save only changed variables.



//Console Settings
cl_chatmode "2"
cl_savehistory "1"
con_clearnotify "0"
con_completion_changed_mark "1"
con_completion_color_changed_mark "f30"
con_completion_color_colon "fff"
con_completion_color_name "8ff"
con_completion_color_quotes_current "ff8"
con_completion_color_quotes_default "ff8"
con_completion_color_title "ff3"
con_completion_color_value_current "fff"
con_completion_color_value_default "fff"
con_completion_format "0"
con_completion_padding "2"
con_funchars_mode "0"
con_hide_chat_input "1"
con_highlight "1"
con_highlight_mark "niomic, nmc"
con_notify "1"
con_notifylines "0"
con_notifytime "15"
con_particles_alpha "1"
con_particles_images ""
con_prompt_charcode "93"
con_shift "10"
con_tilde_mode "0"
con_timestamps "1"
con_wordwrap "0"
cvar_viewdefault "1"
cvar_viewhelp "1"
cvar_viewlatched "1"
gl_alphafont "256"
gl_consolefont "niomic"
gl_smoothfont "1"
log_dir ""
log_readable "1"
scr_conalpha "0.2"
scr_conback "1"
scr_conpicture "conback"
scr_consize "0.4"
scr_conspeed "100000"
scr_printspeed "color"

//FPS and EyeCandy Settings
cl_backpackfilter "0"
cl_deadbodyFilter "2"
cl_fakeshaft "0.65"
cl_fakeshaft_extra_updates "1"
cl_gibFilter "1"
cl_hidenails "0"
cl_hiderockets "0"
cl_independentPhysics "1"
cl_lerp_monsters "0"
cl_maxfps "616"
cl_model_bobbing "0"
cl_muzzleflash "0"
cl_newlerp "0"
cl_nolerp "0"
cl_nolerp_on_entity "0"
cl_novweps "0"
cl_physfps "77"
cl_physfps_spectator "77"
cl_r2g "0"
cl_vsync_lag_fix "0"
cl_vsync_lag_tweak "1.0"
gl_detail "0"
gl_powerupshells "0"
gl_powerupshells_base1level "0.05"
gl_powerupshells_base2level "0.1"
gl_powerupshells_effect1level "0.75"
gl_powerupshells_effect2level "0.4"
gl_powerupshells_size "5"
gl_powerupshells_style "4"
gl_simpleitems "0"
gl_simpleitems_orientation "2"
gl_simpleitems_size "32"
r_bloom "0"
r_bloom_alpha "0.5"
r_bloom_darken "3"
r_bloom_diamond_size "8"
r_bloom_fast_sample "0"
r_bloom_intensity "1"
r_bloom_sample_size "256"
r_chaticons_alpha "0.5"
r_drawentities "1"
r_drawflame "1"
r_drawvweps "1"
r_explosionLight "0"
r_explosionLightColor "5"
r_explosionType "1"
r_flagColor "0"
r_grenadeTrail "7"
r_lerpframes "1"
r_lerpmuzzlehack "1"
r_lgbloodColor "225"
r_lightflicker "1"
r_powerupGlow "2"
r_rocketLight "0"
r_rocketLightColor "7"
r_rocketTrail "7"
r_shaftalpha "1"
r_telesplash "0"

//Particle Effects
gl_bounceparticles "1"
gl_clipparticles "1"
gl_part_blobs "1"
gl_part_blood "0"
gl_part_detpackexplosion_fire_color ""
gl_part_detpackexplosion_ray_color ""
gl_part_explosions "1"
gl_part_gunshots "0"
gl_part_inferno "0"
gl_part_lavasplash "0"
gl_part_spikes "0"
gl_part_telesplash "0"
gl_part_tracer1_color "0 124 0"
gl_part_tracer1_size "3.75"
gl_part_tracer1_time "0.5"
gl_part_tracer2_color "255 77 0"
gl_part_tracer2_size "3.75"
gl_part_tracer2_time "0.5"
gl_part_trails "1"
gl_particle_blobs "0"
gl_particle_blood "0"
gl_particle_blood_color "4"
gl_particle_blood_type "1"
gl_particle_deatheffect "1"
gl_particle_explosions "0"
gl_particle_fasttrails "0"
gl_particle_fire "0"
gl_particle_firecolor ""
gl_particle_gibtrails "0"
gl_particle_gunshots "0"
gl_particle_gunshots_type "1"
gl_particle_muzzleflash "0"
gl_particle_shockwaves "1"
gl_particle_shockwaves_flat "0"
gl_particle_sparks "0"
gl_particle_spikes "0"
gl_particle_spikes_type "1"
gl_particle_style "0"
gl_particle_trail_detail "1"
gl_particle_trail_lenght "1"
gl_particle_trail_time "1"
gl_particle_trail_type "1"
gl_particle_trail_width "3"
gl_solidparticles "0"
gl_squareparticles "0"
r_particles_count "4096"

//Lighting
gl_colorlights "1"
gl_coronas "0"
gl_coronas_tele "0"
gl_fb_bmodels "1"
gl_fb_models "1"
gl_flashblend "1"
gl_lighting_colour "0"
gl_lighting_vertex "0"
gl_lightmode "2"
gl_loadlitfiles "0"
gl_rl_globe "0"
gl_shaftlight "0.5"
r_dynamic "1"
r_fullbright "0"
r_lightmap "1"
r_shadows "0"

//Turbulency and Sky Settings
gl_caustics "0"
gl_fog "0"
gl_fogblue "0"
gl_fogend "800.0"
gl_foggreen "0"
gl_fogred "0"
gl_fogsky "1"
gl_fogstart "400"
gl_surface_lava "0"
gl_surface_slime "0"
gl_turb_trails "0"
gl_turbalpha "1"
gl_turbfog "0"
gl_turbfogDensity "1"
gl_turbfog_color_lava "255 64 0"
gl_turbfog_color_slime "128 255 0"
gl_turbfog_color_water "32 64 128"
gl_turbripple "0"
gl_weather_rain "0"
gl_weather_rain_fast "0"
r_fastsky "0"
r_fastturb "1"
r_lavacolor "0 0 0"
r_novis "0"
r_skycolor "0 0 0"
r_skyname "portaldn.png"
r_slimecolor "0 0 255"
r_telecolor "0 0 0"
r_watercolor "10 10 10"

//Weapon View Model Settings
cl_bob "0"
cl_bobcycle "0.6"
cl_bobup "0"
cl_filterdrawviewmodel "1"
r_drawviewmodel "0"
r_viewmodelSize "0.7"
r_viewmodellastfired "0"
r_viewmodeloffset "0"
r_viewpreselgun "1"

//Texture Settings
gl_anisotropy "0"
gl_ext_arb_texture_non_power_of_two "1"
gl_ext_texture_compression "0"
gl_externalTextures_bmodels "1"
gl_externalTextures_world "1"
gl_lerpimages "1"
gl_lumaTextures "1"
gl_luma_level "1"
gl_max_size "4096"
gl_miptexLevel "0"
gl_no24bit "0"
gl_picmip "0"
gl_playermip "0"
gl_scaleModelTextures "0"
gl_scaleTurbTextures "0"
gl_subdivide_size "128"
gl_textureless "0"
gl_texturemode "gl_linear_mipmap_linear"
gl_texturemode2d "gl_nearest"
r_drawflat "0"
r_floorcolor "30 40 60"
r_wallcolor "10 10 10"

//OpenGL Rendering
gl_affinemodels "1"
gl_buildingsparks "0"
gl_clear "0"
gl_clearColor "0 0 0"
gl_cull "1"
gl_cutf_tesla_effect "0"
gl_detpacklights "0"
gl_extratrails "0"
gl_finish "0"
gl_gammacorrection "0"
gl_inferno_speed "100"
gl_inferno_trail "2"
gl_lightning "0"
gl_lightning_color "120 140 255"
gl_lightning_size "100"
gl_lightning_sparks "0"
gl_lightning_sparks_size "300"
gl_modulate "1"
gl_motiontrails "0"
gl_motiontrails_wtf "0"
gl_nailtrail "0"
gl_nailtrail_plasma "1"
gl_nailtrail_turb "0"
gl_nocolors "0"
gl_smoothmodels "1"
gl_triplebuffer "1"
gl_ztrick "1"
r_farclip "16384"

//System Settings
cl_mediaroot "0"
mumble_debug "0"
mumble_distance_ratio "0.0254"
mumble_enabled "1"
sys_disableWinKeys "0"
sys_highpriority "1"
sys_inactiveSleep "0"
sys_yieldcpu "0"

//Video Settings
_windowed_mouse "1"
vid_allowExtensions "1"
vid_allowSoftwareGL "0"
vid_borderless "0"
vid_colorbits "32"
vid_conaspect ""
vid_conheight "480"
vid_conwidth "853"
vid_customheight "900"
vid_customwidth "1600"
vid_depthbits "0"
vid_displayfrequency "120"
vid_flashonactivity "1"
vid_forcerestoregamma "0"
vid_fullscreen "1"
vid_glDriver "opengl32"
vid_hwgammacontrol "2"
vid_ignoreGLErrors "1"
vid_maskMinidriver "0"
vid_minpos "0"
vid_mode "12"
vid_showextensions "0"
vid_stencilbits "8"
vid_verbose "0"
vid_vsync "0"
vid_wideaspect "0"
vid_xpos "172"
vid_ypos "128"

//Sound Settings
bgmvolume "0.4"
cl_staticSounds "0"
s_ambientfade "100"
s_ambientlevel "0"
s_khz "44"
s_loadas8bit "0"
s_mixahead "0.07"
s_mm1_file "misc/talk.wav"
s_mm1_volume "1"
s_mm2_file "misc/talk.wav"
s_mm2_volume "4"
s_noextraupdate "1"
s_nosound "0"
s_otherchat_file "misc/talk.wav"
s_otherchat_volume "1"
s_precache "1"
s_show "0"
s_spec_file "misc/talk.wav"
s_spec_volume "0"
s_swapstereo "0"
sys_inactiveSound "1"
volume "0.1"



//Network Settings
cl_c2sImpulseBackup "0"
cl_chunksperframe "5"
cl_crypt_rcon "0"
cl_delay_packet "0"
cl_earlypackets "1"
cl_fix_mvd "0"
cl_nodelta "0"
cl_nopred "0"
cl_pext_256packetentities "1"
cl_pext_chunkeddownloads "1"
cl_pext_other "0"
cl_predict_half "0"
cl_predict_players "1"
cl_proxyaddr ""
cl_solid_players "1"
cl_timeout "60"
cl_useproxy "0"
pushlatency "0"

//Status Bar and Scoreboard
cl_hudswap "0"
cl_loadFragfiles "1"
cl_parseFrags "1"
cl_sbar "0"
cl_useimagesinfraglog "0"
con_fragmessages "0"
r_damagestats "0"
scr_centerSbar "1"
scr_compactHud "0"
scr_compactHudAlign "0"
scr_drawHFrags "3"
scr_drawVFrags "0"
scr_newhud "1"
scr_newhud_clear "0"
scr_sbar_drawammo "1"
scr_sbar_drawammocounts "1"
scr_sbar_drawammoicon "1"
scr_sbar_drawarmor "1"
scr_sbar_drawarmoricon "1"
scr_sbar_drawfaceicon "1"
scr_sbar_drawguns "1"
scr_sbar_drawhealth "1"
scr_sbar_drawitems "1"
scr_sbar_drawsigils "1"
scr_scoreboard_borderless "0"
scr_scoreboard_centered "1"
scr_scoreboard_drawtitle "1"
scr_scoreboard_fadescreen "0.5"
scr_scoreboard_fillalpha "0.5"
scr_scoreboard_fillcolored "3"
scr_scoreboard_forcecolors "1"
scr_scoreboard_posy "0"
scr_scoreboard_showfrags "1"
scr_scoreboard_teamsort "0"

//Crosshair Settings
cl_crossx "0"
cl_crossy "0"
crosshair "7"
crosshairalpha "1"
crosshaircolor "0 255 0"
crosshairimage "13_2"
crosshairsize "1.3"

//Screen Settings
cl_clock "0"
cl_clock_format "%H:%M:%S"
cl_clock_x "0"
cl_clock_y "-1"
cl_confirmquit "0"
cl_democlock "0"
cl_democlock_x "16"
cl_democlock_y "-2"
cl_gameclock "3"
cl_gameclock_offset "0"
cl_gameclock_x "28"
cl_gameclock_y "37"
cl_multiview "0"
cl_mvdisplayhud "1"
cl_mvhudflip "0"
cl_mvhudpos "bottom center"
cl_mvhudvertical "0"
cl_mvinset "0"
cl_mvinsetcrosshair "1"
cl_mvinsethud "1"
cl_onload "console"
cl_shownet "0"
cl_verify_qwprotocol "1"
cl_window_caption "1"
demo_playlist_loop "0"
demo_playlist_track_name ""
gl_contrast "1.8"
gl_gamma "0.6"
r_glstats "0"
r_netgraph "1"
r_netstats "1"
r_speeds "0"
r_tracker_align_right "1"
r_tracker_color_bad "900"
r_tracker_color_fragonme "900"
r_tracker_color_good "090"
r_tracker_color_myfrag "099"
r_tracker_color_suicide "900"
r_tracker_color_tkbad "909"
r_tracker_color_tkgood "990"
r_tracker_flags "1"
r_tracker_frags "2"
r_tracker_frame_color "100 100 100 50"
r_tracker_images_scale "1.2"
r_tracker_messages "4"
r_tracker_name_skip_prefix "0"
r_tracker_name_width "0"
r_tracker_own_frag_prefix "You fragged "
r_tracker_postive_enemy_suicide "0"
r_tracker_scale "1"
r_tracker_streaks "0"
r_tracker_string_died " (died)"
r_tracker_string_enemy "enemy"
r_tracker_string_suicides " [SELF]"
r_tracker_string_teammate "teammate"
r_tracker_time "4"
r_tracker_x "0"
r_tracker_y "50"
scr_autoid "0"
scr_autoid_drawname "1"
scr_centershift "10"
scr_centertime "2"
scr_coloredText "1"
scr_coloredfrags "1"
scr_menualpha "0.2"
scr_menudrawhud "1"
scr_qtvbuffer "0"
scr_qtvbuffer_x "0"
scr_qtvbuffer_y "-10"
scr_scoreboard_drawfps "0"
scr_showcrosshair "1"
scr_shownick_frame_color "10 0 0 120"
scr_shownick_name_width "6"
scr_shownick_order "%p%n %a/%H %w"
scr_shownick_scale "1"
scr_shownick_time "0.8"
scr_shownick_x "0"
scr_shownick_y "0"
scr_spectatorMessage "0"
scr_teaminfo "0"
scr_teaminfo_align_right "1"
scr_teaminfo_armor_style "3"
scr_teaminfo_frame_color "0 0 0 100"
scr_teaminfo_loc_width "5"
scr_teaminfo_low_health "10"
scr_teaminfo_name_width "3"
scr_teaminfo_order "%p%n %l‘ %a/%h %w"
scr_teaminfo_scale "0.7"
scr_teaminfo_show_enemies "1"
scr_teaminfo_show_self "2"
scr_teaminfo_weapon_style "0"
scr_teaminfo_x "10"
scr_teaminfo_y "102"
scr_tracking "0"
scr_weaponstats "0"
scr_weaponstats_align_right "1"
scr_weaponstats_frame_color "10 0 0 120"
scr_weaponstats_order "&c990sg&r:%2 &c099ssg&r:%3 &c900rl&r:#7 &c009lg&r:%8"
scr_weaponstats_scale "1"
scr_weaponstats_x "0"
scr_weaponstats_y "0"
show_fps "1"
show_fps_x "-10"
show_fps_y "-10"
show_speed "0"
show_speed_x "-22.5"
show_speed_y "19"
show_velocity_3d "0"
show_velocity_3d_offset_down "5"
show_velocity_3d_offset_forward "2.5"
showdrop "0"
showpackets "0"
showpause "1"
showram "0"
showturtle "0"

//Screen & Powerup Blends
cl_bonusflash "0"
cl_demoplay_flash "1"
gl_cshiftpercent "100"
gl_hwblend "1"
gl_polyblend "0"
v_contentblend "0"
v_damagecshift "0"
v_dlightcshift "0"
v_pentcshift "0"
v_quadcshift "0"
v_ringcshift "0"
v_suitcshift "0"

//View Settings
cl_rollangle "0"
cl_rollspeed "800"
default_fov "120"
fov "133.169998"
r_nearclip "1"
v_centermove "0.15"
v_centerspeed "500"
v_gunkick "0"
v_idlescale "0"
v_ipitch_cycle "1"
v_ipitch_level "0.3"
v_iroll_cycle "0.5"
v_iroll_level "0.1"
v_iyaw_cycle "2"
v_iyaw_level "0.3"
v_kickpitch "0"
v_kickroll "0"
v_kicktime "0"
v_viewheight "-6"
viewsize "110"

//Demo Handling
cl_demoPingInterval "5"
cl_demospeed "1"
cl_startupdemo ""
demo_autotrack "0"
demo_benchmarkdumps "1"
demo_capture_codec "xvid"
demo_capture_dir "movies"
demo_capture_fps "25"
demo_capture_mp3 "0"
demo_capture_mp3_kbps "128"
demo_capture_steadycam "0"
demo_dir "demos"
demo_format "qwd"
demo_getpings "1"
qizmo_dir "demos"
qwdtools_dir "qwdtools"

//Screenshot Settings
image_jpeg_quality_level "100"
image_png_compression_level "1"
scr_allowsnap "1"
sshot_autoname "0"
sshot_dir ""
sshot_format "jpg"

//Spectator Tracking
cam_dist "100"
cam_lockdir "0"
cam_lockpos "0"
cam_thirdperson "0"
cam_zoomaccel "2000"
cam_zoomspeed "300"
cl_camera_death "0"
cl_camera_tpp "0"
cl_camera_tpp_distance "-100"
cl_camera_tpp_height "24"
cl_chasecam "1"
cl_hightrack "0"

//Unsorted Variables
allow_download "1"
allow_download_demos "1"
allow_download_maps "1"
allow_download_models "1"
allow_download_other "1"
allow_download_pakmaps "1"
allow_download_skins "1"
allow_download_sounds "1"
auth_timeout "3600"
auth_validate "1"
auth_warninvalid "0"
cl_curlybraces "0"
cl_warnexec "1"
coop "0"
deathmatch "4"
download_map_url ""
filterban "1"
frag_log_type "0"
fraglimit "0"
fs_cache "1"
gl_maxtmu2 "0"
gl_particle_fulldetail "1"
hostname "unnamed"
in_builtinkeymap "1"
keymap_name "FI/SWE"
maxclients "24"
maxspectators "8"
maxvip_spectators "0"
not_auth_timeout "20"
pausable "1"
pm_airstep "0"
pm_ktjump "1"
qconsole_log_say "0"
qtv_maxstreams "1"
qtv_password ""
qtv_pendingtimeout "5"
qtv_streamport "0"
qtv_streamtimeout "45"
samelevel "64"
skill "3"
spectator_password ""
sv_accelerate "10"
sv_admininfo ""
sv_airaccelerate "10"
sv_allowlastscores "1"
sv_cheats "0"
sv_cpserver "0"
sv_crypt_rcon "1"
sv_cullentities "0"
sv_default_name "unnamed"
sv_demoClearOld "0"
sv_demoDir "demos"
sv_demoExtraNames "0"
sv_demoIdlefps "10"
sv_demoMaxDirSize "102400"
sv_demoMaxSize "20480"
sv_demoPrefix ""
sv_demoRegexp "\.mvd(\.(gz|bz2|rar|zip))?$"
sv_demoSuffix ""
sv_demoUseCache "0"
sv_demofps "30"
sv_demonovis "1"
sv_demopings "3"
sv_demotxt "1"
sv_downloadchunksperframe "2"
sv_enable_cmd_minping "1"
sv_enableprofile "0"
sv_forcenick "0"
sv_forcenqprogs "0"
sv_forcespec_onfull "2"
sv_friction "4"
sv_getrealip "1"
sv_gravity "800"
sv_hashpasswords "1"
sv_kicktop "1"
sv_kickuserinfospamcount "30"
sv_kickuserinfospamtime "3"
sv_ktpro_mode "auto"
sv_loadentfiles "0"
sv_logdir "."
sv_login "0"
sv_mapcheck "1"
sv_maxdownloadrate "0"
sv_maxlogsize "0"
sv_maxrate "0"
sv_maxspeed "320"
sv_maxtic "0.1"
sv_maxuploadsize "1048576"
sv_maxvelocity "2000"
sv_minping "0"
sv_mintic "0"
sv_mod_msg_file ""
sv_nailhack "1"
sv_onDemoRemove ""
sv_onRecordFinish ""
sv_phs "1"
sv_progsname "qwprogs"
sv_progtype "0"
sv_qwfwd_port "30000"
sv_rconlim "10"
sv_reconnectlimit "0"
sv_registrationinfo ""
sv_sayteam_to_spec "1"
sv_serverip ""
sv_specprint "0"
sv_spectalk "1"
sv_spectatormaxspeed "500"
sv_speedcheck "1"
sv_stopspeed "100"
sv_timestamplen "60"
sv_unfake "1"
sv_use_dns "0"
sv_use_internal_cmd_dl "1"
sv_wateraccelerate "10"
sv_waterfriction "4"
sys_restart_on_error "0"
sys_select_timeout "10000"
teamplay "0"
telnet_log_level "0"
telnet_password ""
timelimit "3"
timeout "65"
vip_password ""
vip_values ""
watervis "0"
zombietime "2"
2010-12-24, 03:46
Member
485 posts

Registered:
Feb 2006
Try the death thing online to rule out any local server cpu lag.
2010-12-24, 04:47
Member
284 posts

Registered:
Oct 2006
Found a fix for death fps drop (lulz)

Sure enough it was the scoreboard (like me and others had suspected). And there is a very vicious variable there that causes all the lagging

Set scr_scoreboard_fadescreen "0" and no more death / scoreboard lag I'm glad I found a fix and that that variable's default value is 0, but I'm still shocked that a graphically simple thing like the fading effect would tax your fps by half, incredible. Bad code or what's the deal?

To get rid of menu lagging (mostly)
scr_menualpha "0"

Edit: This only seems to affect the first in-game menu, once you move into Options, FPS goes down dramatically again ;/

And this gave me an idea, because earlier I was figuring I should do scr_conalpha 1, but as it turns out zero is the key.

To get rid of console lagging (mostly)
scr_conalpha "0"

Maybe some of this stuff is very clear to many of you, but I haven't been playing around with these types of settings for quite some time as I haven't had any real need. And of course it helps if the default values are not that fps hungry and all those cfg's that have been passed around might have had better settings.
2010-12-24, 08:19
Member
150 posts

Registered:
Nov 2006
Kalma wrote:
Try the death thing online to rule out any local server cpu lag.

Where can i check that please ?

Niomic:

You have some weird variables that i never saw before, is that from some alpha version ?
cl_nolerp_on_entity
gl_texturemode2d

About your fps:
scr_conpicture "conback", try removing to see if you improve console fps
gl_ztrick "1", i think it will reduce fps using 32-bits, try a timedemo to check it

Other thing, sys_yieldcpu "0", this doesn't cause you mouse/screen unresponsive ?
I have a C2D E6750, if i set CPU Affinity to use only 1 core (cpu o or 1) for ezQuake with sys_yieldcpu 0 seems to work better though.
2010-12-24, 09:26
Member
284 posts

Registered:
Oct 2006
MatriX wrote:
Kalma wrote:
Try the death thing online to rule out any local server cpu lag.

Where can i check that please ?

I guess he just means that I should try it outside of localhost, so just go on a server where you can die (ffa for example) and test it there. But I have already tested it, and it's not a localhost issue.

Quote:
Niomic:

You have some weird variables that i never saw before, is that from some alpha version ?
cl_nolerp_on_entity
gl_texturemode2d

Yes, I think so, quite new to me too.

I use gl_texturemode2d when I'm using different crosshairs.

gl_texturemode2d "gl_linear"
I use this for crosshairs that don't care about smoothing (like locust) or crosshairs that I specifically want smoothed, like the dot you get with crosshair 4 (normally a box).
It also smooths out all the elements in your hud to make it look nicer.

gl_texturemode2d "gl_nearest"
I use this for crosshairs that have to be pixel precise, like some small crosses that I'm using, they look horrible with gl_linear.

Quote:
About your fps:
scr_conpicture "conback", try removing to see if you improve console fps

Console FPS is quite good already with the alphafix I mentioned above.

Quote:
gl_ztrick "1", i think it will reduce fps using 32-bits, try a timedemo to check it

I couldn't confirm any consistent fps gain from either value.

Quote:
Other thing, sys_yieldcpu "0", this doesn't cause you mouse/screen unresponsive ?

Some of the settings in the cfg I pasted do not represent my normal settings, just things for maximum fps. I generally use sys_yieldcpu 1.

Quote:
I have a C2D E6750, if i set CPU Affinity to use only 1 core (cpu o or 1) for ezQuake with sys_yieldcpu 0 seems to work better though.

This simply bottlenecks fps at CPU level for me, no use.
2010-12-24, 09:52
Member
284 posts

Registered:
Oct 2006
At the moment, I know what settings I would need to reach a solid stable 1000 fps, but I'm forgoing that and sticking to some nicer graphics and running at an assured maxfps of 600.

Still need to do some playtesting, but dunno how much room there is to improve without bigger compromises.

There are still a few places in which r_powerupglow 1 + own quad means a decrease in fps from like 2000 to 900 in the same spot vs r_powerupglow 2 :F That's quite sick, especially when that spot is at the path health on dm2 looking towards quad stairs. Likewise there are so many rarely occuring things that outside of those, it would be quite easy to get stable 1500 fps. And some are very map specific.

So now I'm running:

4xFSAA (supersample)
16xFSAF
Threaded optimization
g_weapons
20x184MHz cpu clock
800 MHz gpu clock
2000 MHz vidmem

And these confs:

//Console Settings
con_clearnotify "0"
con_completion_format "1"
con_highlight "1"
con_highlight_mark "niomic, nmc"
con_notifylines "0"
con_notifytime "15"
con_particles_images ""
con_timestamps "1"
gl_consolefont "niomic"
scr_conalpha "0"
scr_consize "0.4"
scr_conspeed "100000"

//FPS and EyeCandy Settings
cl_deadbodyFilter "2"
cl_fakeshaft "0.65"
cl_gibFilter "1"
cl_maxfps "600"
cl_model_bobbing "0"
cl_muzzleflash "0"
cl_physfps "77"
cl_physfps_spectator "77"
gl_powerupshells "0"
r_chaticons_alpha "0.5"
r_explosionLight "0"
r_explosionType "3"
r_grenadeTrail "7"
r_rocketLight "0"
r_rocketTrail "7"
r_telesplash "0"

//Particle Effects
gl_part_tracer2_color "0 255 0"
gl_part_tracer2_time "0.4"
gl_part_trails "1"
r_particles_count "1024"

//Lighting
gl_flashblend "1"
gl_shaftlight "0"

//Turbulency and Sky Settings
r_fastsky "1"
r_fastturb "1"
r_lavacolor "0 0 0"
r_skycolor "0 0 0"
r_slimecolor "0 0 255"
r_telecolor "0 0 0"
r_watercolor "10 20 30"

//Weapon View Model Settings
cl_filterdrawviewmodel "1"
r_drawviewmodel "0"
r_viewmodelSize "0.7"
r_viewpreselgun "1"

//Texture Settings
gl_anisotropy "16"
gl_max_size "4096"
gl_texturemode "gl_linear_mipmap_nearest"
gl_texturemode2d "gl_nearest"

//Config Management
cfg_legacy_exec "0"

//OpenGL Rendering
gl_ztrick "1"

//System Settings
sys_highpriority "1"
sys_yieldcpu "1"

//Video Settings
vid_colorbits "32"
vid_conheight "480"
vid_conwidth "853"
vid_displayfrequency "120"
vid_mode "12"

//Sound Settings
cl_staticSounds "0"
s_ambientlevel "0"
s_khz "44"
s_mixahead "0.07"
s_mm2_volume "4"
s_spec_volume "0"
sys_inactiveSound "1"
volume "0.1"

//Input - Keyboard
cl_upspeed "500"
cl_weaponhide "1"
cl_weaponpreselect "1"

//Input - Mouse
m_forward "0"
m_pitch "-0.044"
m_showrate "1"
m_side "0"
m_yaw "0.044"

//Input - Misc
sensitivity "0.4"

//Network Settings
cl_timeout "20"

//Server Browser
sb_autoupdate "0"
sb_findroutes "1"
sb_listcache "1"

//Status Bar and Scoreboard
con_fragmessages "0"
scr_newhud "1"
scr_scoreboard_fillalpha "0.5"
scr_scoreboard_fillcolored "3"
scr_scoreboard_teamsort "0"

//Crosshair Settings
crosshair "7"
crosshaircolor "0 255 0"
crosshairimage "13_2"
crosshairsize "1.3"

//Screen Settings
cl_democlock_x "16"
cl_gameclock "3"
cl_gameclock_x "28"
cl_gameclock_y "37"
cl_onload "console"
gl_contrast "1.8"
gl_gamma "0.6"
r_netgraph "1"
r_netstats "1"
r_tracker_color_myfrag "099"
r_tracker_color_tkbad "909"
r_tracker_flags "1"
r_tracker_frags "2"
r_tracker_frame_color "100 100 100 50"
r_tracker_images_scale "1.2"
r_tracker_messages "4"
r_tracker_string_suicides " [SELF]"
r_tracker_y "50"
scr_autoid "0"
scr_centershift "10"
scr_coloredfrags "1"
scr_menualpha "0"
scr_menudrawhud "1"
scr_spectatorMessage "0"
scr_teaminfo "0"
scr_teaminfo_frame_color "0 0 0 100"
scr_teaminfo_low_health "10"
scr_teaminfo_name_width "3"
scr_teaminfo_order "%p%n %l‘ %a/%h %w"
scr_teaminfo_scale "0.7"
scr_teaminfo_show_enemies "1"
scr_teaminfo_x "10"
scr_teaminfo_y "102"
scr_tracking "0"
scr_weaponstats "0"

//Screen & Powerup Blends
gl_polyblend "0"
v_contentblend "0"

//View Settings
fov "133.169998"
v_kicktime "0"
v_viewheight "-6"
viewsize "110"

//Match Tools
match_auto_logconsole "0"

//Demo Handling
demo_autotrack "1"
demo_capture_codec "xvid"
demo_capture_dir "movies"
demo_capture_fps "25"
demo_dir "demos"
demo_format "qwd"
qizmo_dir "demos"

//Item Names
tp_name_rl "{&cf13rl&cfff}"
tp_name_lg "{&c2aalg&cfff}"
tp_name_ga "{&c0b0ga&cfff}"
tp_name_ya "{&cff0ya&cfff}"
tp_name_ra "{&cf00ra&cfff}"
tp_name_armortype_ga "{&c0b0g&cfff}"
tp_name_armortype_ya "{&cff0y&cfff}"
tp_name_armortype_ra "{&cf00r&cfff}"
tp_name_rockets "rox"
tp_name_quad "{&c05fquad&cfff}"
tp_name_pent "{&cf00pent&cfff}"
tp_name_ring "{&cff0ring&cfff}"
tp_name_mh "{&c0a0mega&cfff}"
tp_name_health "{&c0b0hth&cfff}"
tp_name_armor "{&cfa0arm&cfff}"
tp_name_weapon "{&cc90wep&cfff}"
tp_name_backpack "{&cf2apack&cfff}"
tp_name_enemy "{&cf00enemy&cfff}"
tp_name_quaded "{&c05fquaded&cfff}"
tp_name_pented "{&cf00pented&cfff}"
tp_name_rune4 "{rgn}"
tp_name_rune3 "{hst}"
tp_name_rune2 "{str}"
tp_name_rune1 "{res}"
tp_name_someplace " "
tp_name_at " "

//Item Need Amounts
tp_need_weapon "87"
tp_need_ga "70"
tp_need_ya "70"
tp_need_ra "70"
tp_need_shells "0"
tp_need_nails "0"
tp_need_cells "13"
tp_need_health "40"

//Unsorted Variables
deathmatch "4"
keymap_name "FI/SWE"
pausable "1"
pm_airstep "0"
skill "3"
timelimit "3"
2010-12-24, 10:12
Moderator
1329 posts

Registered:
Apr 2006
gl_texturemode "gl_linear_mipmap_nearest"

What's the point? You are doing anisotropic filtering but mipmap transitions are bilinear? That's horrible, use gl_linear_mipmap_linear.


gl_max_size "4096"

Remove this one, it will be autodetected from the driver after that (it should default to 16384 with gtx400 cards).


sys_highpriority "1"
sys_yieldcpu "1"

Yieldcpu allows background applications to have CPU-time, yet you specify highpriority 1? Do not use highpriority 1 unless you have serious problems with something like input/fps drops NOT related to ezQuake itself.


cl_maxfps "600"

Use 616 instead to make sure ezQuake internal timings are in sync with physfps. Yes, it still has an effect to everything.


gl_ztrick "1"

You will lose fps but gain nothing. Use 0 here.


Also as I don't see gl_tripplebuffer variable in the config, I'm guessing it's defaulted to 1, change it to 0. I've seen it do harm and lower fps instead of being useful.
Servers: Troopers
2010-12-24, 10:40
Member
284 posts

Registered:
Oct 2006
Funny thing is that most of the things you commented on where like you suggested, _before_ I started messing around with my config

Renzo wrote:
gl_texturemode "gl_linear_mipmap_nearest"

What's the point? You are doing anisotropic filtering but mipmap transitions are bilinear? That's horrible, use gl_linear_mipmap_linear.

I changed to "gl_linear_mipmap_nearest" after I noticed that it gives better fps (70fps in timedemo and standing at dm3 quad ledge) and I could honestly not notice a difference that would matter.
Is there some specific situation where I could notice it better or what is the reasoning. I don't feel like wasting ~70fps on something I can't see

Based on this, I also went back to 2xFSAA and only multisampling. Can't really see enough difference, especially when playing. I'm also quite content with how 4xFSAF looks, but will keep it at 16x for now, as there I can definitely see the difference and the FPS hit seems to be quite small, considering the difference.

Renzo wrote:
gl_max_size "4096"

Remove this one, it will be autodetected from the driver after that (it should default to 16384 with gtx400 cards).

Removed.

Renzo wrote:
sys_highpriority "1"
sys_yieldcpu "1"

Yieldcpu allows background applications to have CPU-time, yet you specify highpriority 1? Do not use highpriority 1 unless you have serious problems with something like input/fps drops NOT related to ezQuake itself.

Highpriority was just for testing now, haven't really noticed any difference so might as well switch back to 0.
From what I understand, I should still leave yieldcpu on?

Renzo wrote:
cl_maxfps "600"

Use 616 instead to make sure ezQuake internal timings are in sync with physfps. Yes, it still has an effect to everything.

This was simply based on 5*120 as I'm using earlypackets, I thought that engine fps syncing didn't matter. Switched to 616 even before reading this, cos that's what I had before


Renzo wrote:
gl_ztrick "1"

You will lose fps but gain nothing. Use 0 here.

Well, two people have now said that I shouldn't have it as 1, switched to 0

Renzo wrote:
Also as I don't see gl_tripplebuffer variable in the config, I'm guessing it's defaulted to 1, change it to 0. I've seen it do harm and lower fps instead of being useful.

Changed to 0
2010-12-24, 10:52
Member
462 posts

Registered:
Jan 2006
Renzo wrote:
gl_texturemode "gl_linear_mipmap_nearest"

What's the point? You are doing anisotropic filtering but mipmap transitions are bilinear? That's horrible, use gl_linear_mipmap_linear.

At least for me linear_mipmap_nearest gives 300fps more compared to linear_mipmap_linear, that's why I use it.
2010-12-24, 10:59
Member
284 posts

Registered:
Oct 2006
blAze wrote:
Renzo wrote:
gl_texturemode "gl_linear_mipmap_nearest"

What's the point? You are doing anisotropic filtering but mipmap transitions are bilinear? That's horrible, use gl_linear_mipmap_linear.

At least for me linear_mipmap_nearest gives 300fps more compared to linear_mipmap_linear, that's why I use it.

Can you tell the difference, and if so, where ? And I'm talking about differences that you could actually see when playing, not just standing still and staring at the screen
2010-12-24, 10:59
Moderator
1329 posts

Registered:
Apr 2006
niomic wrote:
I changed to "gl_linear_mipmap_nearest" after I noticed that it gives better fps (70fps in timedemo and standing at dm3 quad ledge) and I could honestly not notice a difference that would matter. Is there some specific situation where I could notice it better or what is the reasoning. I don't feel like wasting ~70fps on something I can't see

Mipmap transitions are not smoothed out when bilinear is used instead of trilinear. With the usage of anisotropic filtering the "line" is just pushed further away but it is still there. It is especially visible when you move, try to see it first without anisotropic filtering and you know what to look, then turn on the anisotropic filtering.

There's your difference:

http://img98.imageshack.us/img98/6135/79631757.jpg



niomic wrote:
From what I understand, I should still leave yieldcpu on?

You can leave it enabled, I don't think there's anything that can hog the CPU resources so badly that it would affect your client. I had to run prime95 with 8 threads to actually see a fps drop (which is significant in that case). In that case sys_yieldcpu 0 helped a lot since prime95 is a "nice" process (if required, it gives out the CPU resources it uses).
Servers: Troopers
2010-12-24, 11:04
Member
462 posts

Registered:
Jan 2006
niomic wrote:
blAze wrote:
At least for me linear_mipmap_nearest gives 300fps more compared to linear_mipmap_linear, that's why I use it.

Can you tell the difference, and if so, where ? And I'm talking about differences that you could actually see when playing, not just standing still and staring at the screen

I haven't done any extensive testing, but doing a /timerefresh on the dm3 ring having the ring, mega and quad visible, it's like 2300 vs 2600 fps.

Or did you mean visual diff? Don't know, don't care.
2010-12-24, 11:21
Member
284 posts

Registered:
Oct 2006
Renzo wrote:
niomic wrote:
I changed to "gl_linear_mipmap_nearest" after I noticed that it gives better fps (70fps in timedemo and standing at dm3 quad ledge) and I could honestly not notice a difference that would matter. Is there some specific situation where I could notice it better or what is the reasoning. I don't feel like wasting ~70fps on something I can't see

Mipmap transitions are not smoothed out when bilinear is used instead of trilinear. With the usage of anisotropic filtering the "line" is just pushed further away but it is still there. It is especially visible when you move, try to see it first without anisotropic filtering and you know what to look, then turn on the anisotropic filtering.

There's your difference:

http://img98.imageshack.us/img98/6135/79631757.jpg

I don't have that kind of problem, but switching from nearest to linear in that exact spot causes about a 130fps drop.

In nvidia settings, I have:
- Ambient Occlusion: Off
- Anisotropic filtering: Application-controlled
- Antialiasing - Gamma correction: Off
- Antialiasing - Mode: Override any application setting
- Antialiasing - Setting: 2x
- Antialiasing - Transparency: Multisample
- CUDA - GPU's: All
- Extension limit: Off
- Maximum pre-rendered frames: 3
- Multi-display/mixed-gpu acceleration: Single display performance mode
- Power management mode: Prefer maximum performance
- Preferred refresh rate (BenQ XL2410T): Application-controlled
- Texture filtering - Anisotropic sample optimization: On
- Texture filtering - Negative LOD bias: Allow
- Texture filtering - Quality: Quality
- Texture filtering - Trilinear optimization: On
- Threaded optimization: On
- Triple buffering: Off
- Vertical sync: Use the 3D application setting



This shot is taken with nearest, and I can't see any real difference switching from one to the other, with or without vid_restart, although I assume it's not needed as I can see the FPS drop immediately.

http://i208.photobucket.com/albums/bb83/cimoin/ezquake000-1.png
2010-12-24, 11:23
Member
284 posts

Registered:
Oct 2006
blAze wrote:
niomic wrote:
blAze wrote:
At least for me linear_mipmap_nearest gives 300fps more compared to linear_mipmap_linear, that's why I use it.

Can you tell the difference, and if so, where ? And I'm talking about differences that you could actually see when playing, not just standing still and staring at the screen

I haven't done any extensive testing, but doing a /timerefresh on the dm3 ring having the ring, mega and quad visible, it's like 2300 vs 2600 fps.

Or did you mean visual diff? Don't know, don't care.

Visual difference indeed, because as I said, I'm experiencing an fps drop and can't even see any difference. If I could see a difference, then it would be a prioritizing question
2010-12-24, 11:39
Moderator
1329 posts

Registered:
Apr 2006
niomic wrote:
- Power management mode: Prefer maximum performance
- Texture filtering - Negative LOD bias: Allow
- Vertical sync: Use the 3D application setting

This shot is taken with nearest, and I can't see any real difference switching from one to the other, with or without vid_restart, although I assume it's not needed as I can see the FPS drop immediately.

You should use:

1) adaptive (it doesn't affect your fps at all, the card downclocks itself if the game is not stressful enough)
2) clamp (makes anistropic filtering more constant on different games / reduces texture shimmer)
3) force off (why use vsync in the first place?)


Also, why did you use the same spot as I did? It was only to show how bilinear/trilinear filtering differ and I already said the transition is pushed back/further away by anisotropic filtering.


Do this, go to the low-rl (near ga/tele) and look towards ya+100h. Move forward and back (don't go the stairs down, stay near the rl) and look at the right side of the wall. Change between gl_linear_mipmap_linear and gl_linear_mipmap_nearest and repeat it few times. The wall on the left looks like it's shimmering when bilinear is used, it's better when trilinear is used even if the difference is not that great tbh, but even so, the difference is there.

Again, that is only one situation on one map, there are bigger areas in QW maps, like in dm3 for example, or dm2 corridors. However this seems to be highly texture-dependant, so it probably won't be visible on most of the textures.
Servers: Troopers
2010-12-24, 11:50
Member
284 posts

Registered:
Oct 2006
Renzo wrote:
niomic wrote:
- Power management mode: Prefer maximum performance
- Texture filtering - Negative LOD bias: Allow
- Vertical sync: Use the 3D application setting

This shot is taken with nearest, and I can't see any real difference switching from one to the other, with or without vid_restart, although I assume it's not needed as I can see the FPS drop immediately.

Why do you use the same spot as I did? It was only to show how bilinear/trilinear filtering differ and I already said the transition is pushed back/further away by anisotropic filtering.


Do this, go to the low-rl (near ga/tele) and look towards ya+100h. Move forward and back (don't go the stairs down, stay near the rl) and look at the right side of the wall. Change between gl_linear_mipmap_linear and gl_linear_mipmap_nearest and repeat it few times. The wall on the left looks like it's shimmering when bilinear is used, it's better when trilinear is used even if the difference is not that great tbh, but even so, the difference is there.

Again, that is only one situation on one map, there are bigger areas in QW maps, like in dm3 for example, or dm2 corridors.

I believe it is better, as it is the best choice available in qw and the most resource hungry. I tested that situation and did notice some shimmering in the walls near where the textures connect and there's a graphical tile outline in the texture pictures. Funny thing is that in this situation, the fps difference is negligible :]] But with both nearest and linear, I noticed shimmering in the ceiling.

I'll try some games with linear on again, but if it's causing me to go under 616, I will gladly go back to nearest as there are so few situations where I would notice it

Thanks for taking the time to make the point anyway.

And I made the changes you suggested in nvidia drivers. I haven't gone through them one by one, so that's why vsync was still optional (it's disabled in qw anyway).

Now I'm off till Sunday, so no more testing Happy holidays!
2010-12-24, 13:00
Administrator
384 posts

Registered:
Dec 2006
niomic wrote:
I would assume everyone gets a noticeable fps drop from g_weapons and especially dying, doesn't mean the issue shouldn't be addressed? I agree that the console fps drop isn't that bad, but it's not even close in magnitude to the drop you get when you die. And there the problem is that it takes a little time for qw to get back up-to-speed after you respawn, as in it is noticeable when I play, if I don't play with a low enough cl_maxfps to mitigate even that problem and that would mean that I'm running at 3/4 or 3/5 of what I could get stable without this dying fps drop.

I deliberately didn't say anything about g_weapons or dying because I perceived that they are things that could affect the game if you get fps drops.

BTW a good place I've found for testing performance is if you stand on quad on dark-terror-ffa.bsp (it's a map on xs4all ffa). Easily the slowest I've seen ezquake-gl run and I probably need to cap at 616fps there.

Good spot on the r_particles_count thingy, gives me around 5-10% extra fps.
2010-12-24, 13:23
Moderator
1329 posts

Registered:
Apr 2006
Heh, I actually did some digging and found something very interesting in DM3:

http://img59.imageshack.us/img59/1958/49963449.jpg


In that specific spot, the performance difference between trilinear filtering vs 16x anisotropic filtering is "only" 136,6% (if you are bad with numbers, it means over two times faster as you can see from the fps)


EDIT:

I'm using "high quality" in the NVCPL under texture filtering quality. Setting that to "quality" and enabling both trilinear and anisotropic filtering optimization (set to: on) increases the fps up to 980fps at lowest.


EDIT 2:

And ah, the bug in ezQuake. If you have set gl_ext_texture_compression 1 and think it's actually enabled, you are wrong (the bug has been there for a LONG time). If I use that spot, set it to 0, do vid_restart, set it back to 1 and do vid_restart, the fps goes above 1000 even with "high quality" filtering and 16x anisotropy.

(texture compression improves performance)
Servers: Troopers
2010-12-24, 15:13
Member
462 posts

Registered:
Jan 2006
gl_ext_texture_compression 1 completely fucks up my crosshairs?! It does give ~150 fps boost though, but the corsshairs should remain the same for it to be usable. r_particles_count did nothing for me.
2010-12-24, 15:29
Moderator
1329 posts

Registered:
Apr 2006
blAze wrote:
gl_ext_texture_compression 1 completely fucks up my crosshairs?!

It shouldn't. Maybe something to do with gl_texturemode and possibly gl_texturemode2d if you use some ezQuake alpha-build. I also remember anisotropy affecting crosshair textures for some reason.
Servers: Troopers
2010-12-24, 15:40
Member
462 posts

Registered:
Jan 2006
I'm using just the plain old 2.0. Looks like the compression minimizes the hit from mipmap_linear making it almost as fast as mipmap_nearest. So if the crosshair problem gets fixed I guess the way to go is mipmap_linear and compression 1.
2010-12-25, 17:52
Member
150 posts

Registered:
Nov 2006
niomic wrote:
MatriX wrote:
I have a C2D E6750, if i set CPU Affinity to use only 1 core (cpu o or 1) for ezQuake with sys_yieldcpu 0 seems to work better though.

This simply bottlenecks fps at CPU level for me, no use.

If im not mistaken, ezQuake does not use more than 1 core, so you will not bottleneck cpu because it will make no diference.
In my case i don't know why, but ezQuake seems to respond better when i set cpu affinity for cpu 1 with sys_yieldcpu 1.
2010-12-25, 18:38
Moderator
1329 posts

Registered:
Apr 2006
MatriX wrote:
If im not mistaken, ezQuake does not use more than 1 core, so you will not bottleneck cpu because it will make no diference. In my case i don't know why, but ezQuake seems to respond better when i set cpu affinity for cpu 1 with sys_yieldcpu 1.

Maybe time for a re-check?

Set affinity for one core only, when multithreaded optimization is enabled, and lose 10% of the performance.
Servers: Troopers
2010-12-25, 19:12
Member
150 posts

Registered:
Nov 2006
You mean Threaded Optimization in Nvidia Control Panel?
Forgot about that since i have it disabled. True.
  55 posts on 2 pages  First page12Last page