• Mitspieler gesucht? Du willst dich locker mit der Community austauschen? Schau gerne auf unserem ComputerBase Discord vorbei!

Total War Community Talk: Rome II - Total War

Sublogics schrieb:
@AlQamar, ich würde auch behaupten daß das Spiel nicht CPU-limitiert ist. Das Limit liegt dann irgendwo zwischen den 1,9 und 3,8 GHz. Und darüber skaliert es zu 100% auf GPU.

ich verstehe deine Aussage nicht. Es ist nicht CPU Limitiert und das Limit liegt bei der CPU Geschwindigkeit?

punkrockfan schrieb:
Und wie das Spiel CPU limitiert ist, die frames der Karten aus dem pcgameshardware test erreichtst Du mit nem Quadcore unterhalb des 83xx von AMD im Leben nicht, die Minimum schon gar nicht.

Kasmopaya schrieb:

ich hab keine GTX 970

Sublogics schrieb:
Aber du siehst doch das ab einem bestimmten Takt keinerlei Steigerung mehr erfolgt. Ab dem Punkt ist dann das CPU-Limit überschritten. Es hängt dann woanders.
Woran liegt es denn, dass es je nach Grafikeinstellung skaliert und dann wieder nicht. Das ist mir nicht klar.
 
Wie ist denn eure Meinung zu den Zahlen?
 
Zu den Werten bei 1920x1080: Im Vergleich zu Standardtakt bringt eine Erhöhung des CPU-Taktes null FPS. Das bedeutet daß kein CPU-Limit vorhanden ist. Erst wenn die CPU stark untertaktet wurde fangen die FPS an zu sinken, dann erst limitiert wieder die CPU.
Und die FPS verhalten sich in meinen Augen bei den unterschiedlichen Grafikeinstellungen so wie sie sollen. Ich sehe nicht was daran keinen Sinn ergeben soll.

Bei der viel geringeren Auflösung von 1024x768 rückt man zudem wieder näher an eine CPU-Limitierung, deshalb spielt da der Takt auch wieder mehr Einfluss.
Sieht für mich alles korrekt aus.

Ich hab allerdings nicht die Werte unter "performance scalation cpu" etc. verglichen. Was sollen die einem sagen?
 
Sublogics schrieb:
Zu den Werten bei 1920x1080: Im Vergleich zu Standardtakt bringt eine Erhöhung des CPU-Taktes null FPS. Das bedeutet daß kein CPU-Limit vorhanden ist. Erst wenn die CPU stark untertaktet wurde fangen die FPS an zu sinken, dann erst limitiert wieder die CPU.
Und die FPS verhalten sich in meinen Augen bei den unterschiedlichen Grafikeinstellungen so wie sie sollen. Ich sehe nicht was daran keinen Sinn ergeben soll.

Bei der viel geringeren Auflösung von 1024x768 rückt man zudem wieder näher an eine CPU-Limitierung, deshalb spielt da der Takt auch wieder mehr Einfluss.
Sieht für mich alles korrekt aus.

Ich hab allerdings nicht die Werte unter "performance scalation cpu" etc. verglichen. Was sollen die einem sagen?

Es gibt jemanden der für mich mit einem einem 770 SLI die Tests wiederholt hat, und dort skaliert sein Spiel linear, also optimal, da offenbar keine GPU Limitierung existiert. D.h. 18% mehr CPU Takt = 18% mehr FPS.

Das finde ich komisch, denn im Umkehrschluss würde die fehlende oder inkonsistente Skalierung auf meinem System bedeuten, dass es bei fehlender Skalierung ein GPU Limit gibt, sogar bei nur 1024er Auflösung.

Was überdies auffällig ist, ist das die Werte mit 1024er Auflösung so ungleichmäßig sind. Mal skalieren sie etwas, mal überhaupt nicht, mal negativ und manchmal sogar über dem Soll.

Glossar:
performance scalation cpu clock: ist der Leistungszuwachs / Multiplikator im Vergleich zum Basistakt (default clock).
1.5 entsprechen 50% mehr Leistung, 0.5 -50% Leistung etc., 1.34 also + 34% Leistung die zu erwarten wären.

performance scalation percent: die tatsächliche Skalierung im Vergleich zum Basistakt.
Bsp: hypothetical wäre 1.34 also 34%, performance scalation percent beträgt 103, haben wir eine Abweichung der Leistung von einer theoretischen Skalierung von 31%. D.h. die Framerate ist deutlich unter dem Niveau den man für eine 1:1 also optimale Skalierung erwarten könnte.

hypothetical FPS: Ausgehend vom Basistakt * performance scalation cpu clock ergibt die rechnerisch zu erwartende Framerate

S17uId.png
 
Zuletzt bearbeitet:
Naja, daß es bei einem SLI aus zwei 770ern wieder auf CPU skaliert wundert mich nicht sooo sehr. Anscheinend ist die 670er nur schon ganz allmählich zu schwach für Attila / Rome 2. Beziehungsweise anders gesagt: Attila scheint recht ausgewogen zu sein was die CPU- und GPU-Nutzung angeht, wenn man es relativ einfach schafft in beide Limits zu geraten.
 
Tritratrullalla.....DER 1. PATCH IST DA

1,3 GB am Laden, mal schauen was dann wie läuft.
 
Uiuiui... na dann bin ich mal gespannt =)

Wird der DLC eigentlich auch bereits heute ausgeliefert? Im Steam Shop ist der nämlich noch für Morgen datiert.
 
Gibt es patch notes? Könnte man dier hier im Forum posten? Die Firewall hier in der Firma ist sehr streng ;)
 
Leider nur auf Englisch gefunden. Als Datum wird zwar der 26.02. angegeben, aber das Update wurde ja um eine Woche verschoben:

Full update notes

Update coming Thursday 26/02/2015

Gates
• Torching settlement gates has been replaced with the ability to hack them down with melee attacks. This new last-resort feature has been made in response to community feedback, and is an important mechanic to enable the attacking side to still have a chance of besieging a settlement even when their siege engines have been destroyed. The trade-off is that it comes at an extremely high price for the attackers. It’s the pyrrhic act of a desperate general. ◦ Certain weapons are more effective at destroying gates than others. Axes and two-handed weapons are the most effective, followed by swords and maces, followed by spears, daggers and pike. The less effective a weapon is, the more time it takes to break through.
◦ Given the prominence of towers, boiling oil etc in ATTILA, attacking gates is a very costly exercise for the attacker, and the AI will only use it as a last resort.
◦ If the AI attacks gates, it will employ more heavily armoured troops to do so, though it will deprioritise the use of elephants and its general.
◦ Basic barbarian settlement gates can be burned down at range using fire-arrows, though this is less efficient than before. Reinforced or bronze gates are immune to fire altogether, and when no artillery is present, will need to be engaged in melee.
◦ The AI is more likely to build siege equipment on the campaign map before assaulting a settlement.


Technical and Performance improvements

Battle
• Level of detail on certain units in battle is now handled better on Mac.
• Improved how the level of detail changes on trees on the campaign map, to make it less noticeable.
• Optimisation for unit banners in battles.
• Memory optimisation for imposter sprites.
• Fixed various instances which may cause crashes.
• Added missing Hadrian’s Wall back into battle outfields.

Campaign
• Fixed a 30-60 second freeze that sometimes occurred after ending turn in Campaign modes.
• Improved video playback.
• Fix for the Campaign map terrain sometimes turning dark/black after leaving a Campaign battle during the End Turn sequence on Mac.
• Some beaches around the coast of Britain will no longer appear white through the Campaign shroud, as they did before with certain graphics settings.
• End Turn times optimised.
• Fix for error in event generation logic.
• Fixed various instances which may cause crashes.

General
• Improved performance with AMD graphics cards.
• AMD MSAA optimisation.
• Changing the "Unit Detail" graphics settings during battle will no longer cause some units to lose their weapons.
• Optimisation to reduce the VRAM usage in battles.

Gameplay Improvements

Battle
• Updated some battle environments.
• Updated Barbarian gate house, shifting the gate forward so attacking units would be effected by boiling oil.
• Fix for a battle map near Constantinopolis, where the terrain previously allowed the player to deploy land units in midair above the sea
• AI controlled elephants are now less likely to attack settlement gates in battles.
• Improved the coordination of the siege AI when attacking walled settlements in order to use their artillery and siege equipment more effectively.
• Added more variation to the environment lighting in late game battles.
• Campaign AI no longer always accepts war coordination targets.
• Improved Battle AIs placement of units within settlements.
• Fixed a battle AI bug which caused units to reform unnecessarily when being commanded by the AI to attack walls.

Campaign
• Updates to victory conditions: ◦ Reduced number of units required for Western Roman Empire to achieve Divine Victory.
◦ Reduced number of buildings required for Divine Victory for Western and Eastern Roman Empires.
◦ Fix for the wrong port building being specified in the Divine Triumph Victory requirements for the Eastern Roman Empire.

• Improved Prologue scripting to prevent some strange behaviours, such as the Huns wandering off in the late stages of the prologue, the Eastern Roman Empire not declaring war for many turns at the end of the prologue, and a progression blocking issue when the player did not own Apulum at the time the script expected them to.
• Campaign AI is now more likely to hold sieges for longer.
• Campaign AI is less likely to underestimate its chances when attacking enemies in Forced March stance.
• "Controlled water" morale effects are now also applied in "Shared water" in the sea during Campaigns.
• Fix for armies controlled by Campaign AI sometimes getting stuck in a settlement’s Zone of Control.
• Fix for a bug causing Blockades and Sieges to not end, when a settlement was razed during Campaign modes.
• Non-Horde region-less Factions, and Hordes with navies but no armies, are now forced to capture settlements instead of razing, sacking etc.
• Fix for "Governed Province" bonuses applying as "Local Province" bonuses when the character is not a Governor in Campaign modes.
• Infighting effects are now removed when going to sea in Campaign modes.
• Infighting effects are now removed when a faction is no longer a Horde in Campaign modes.
• Fixed some campaign Dilemmas that were not spawning the agents they promised.
• The "Employ Him" option on the Stonemason subject dilemma no longer kills a random character.
• The "Do Nothing" option on the Tribal Loyalty" subject dilemma no longer has a public order reduction effect.
• Fixing issue where Huns would not get any missions after a prolonged period of play.
• Added a rematch button for final battle in the Prologue when the player loses.
• Adding more conditions to diplomatic events in Campaign modes.
• The calculation of the "Ferocious Warriors" passive bonus has been changed to better match its tooltip. It was previously calculated based on casualties sustained during the current battle. It is now calculated based on casualties compared to the unit’s full strength.

Balancing Changes

Battle
• Balance pass on building damage.
• Fixed shields appearing too dark on the battlefield.
• Corrected some unit selection voiceovers.
• Music-ducking when horses pass under the camera has been improved.
• Burning embers on trees in battle have been improved on Mac.
• Fixed a descynchronisation in the Multiplayer battle setup lobby, which could cause a crash.
• Removed "Fire damage" from gates tool-tip in battles, as they can no longer be burnt down.
• Refined wavering animations and fixed additive animation on certain units
• Vegetation now correctly impedes projectiles.
• Added a new dynamic state to the music to break up the intensity in battles.
• Fixed text clipping on second player’s faction dropdown in the battle setup screen.
• Fixed the rotation of certain upgraded battle tiles, so they fit into the terrain properly.
• Fixed incorrect positioning/scaling of unit marks in battle, which sometimes obscured the unit health bars on unit markers when the games resolution was set below native resolution. Also improved the sharpness of these when the game is played above native resolution.
• The sun colour in the battle environment no longer affects the colour of the battle deployment areas.
• Fixed clipping and visibility issues with some Longbow, Camel and Falx units.
• Unit markers now appear in the Tactical Map view during battles, when enabled with the "Show unit markers" game setting.
• Decals on top of settlement walls in battles will "smear" less down the face of the wall.

Campaign
• Rebalanced spawned Hunnic enemy armies in Campaign Modes.
• Various subjects and their effects have been updated.
• Fixed a texture on the nomadic spy model in Campaign modes.
• Fix for lighting with climate change in Campaign modes.
• Fixed a screen corruption for the Player 2, which occurred while Player 1 watched the "Attila is born" full-screen cinematic in Multiplayer Campaign mode.
• The correct attribute tooltip is now displayed in the factors tooltip when hovering the mouse over attribute icons in the Diplomacy panel.
• Defeating an army that is besieging an ally's settlement will no longer cause the occupy settlement UI to appear in Campaign Modes.
• Public order mission will no longer remain after being completed in the Prologue.
• When researching the "Manoral Lands Civil" technology in the Prologue, the exclamation mark to indicate that the technology is fully researched will no longer appear 1 turn before it is completed.
• Fixed cases where the player could not allocate a General into the Warlord seat in Office after meeting the requirements in Campaign modes.
• Characters will no longer disappear from the Family Tree when married off to a Widow.
• Players are no longer able to construct multiple lower-level variants of the same building in hordes when this is not the intended design.
• Migrating Barbarian civic buildings no longer transform into an industrial building when demolished.
• Added missing building entries to the encyclopaedia.
• Changes to how the Campaign Victory screen is displayed when completing multiple Victory conditions: ◦ For every victory type, the victory panel with "Quit" and "Continue" options is displayed.
◦ The first ("Minor") victory does not play a movie, just the panel.
◦ The second ("Cultural or Military") will show a movie, then the Victory panel.
◦ After choosing to "Continue", achieving the other ("Divine Triumph" and the remaining "Cultural" or "Military") Victory conditions will once more prompt the player with the Victory panel with options to "Quit" or "Continue".

• When clicking the faction icon for a faction with no regions owned and no faction leader in the Diplomacy panel, the camera will now centre on an army or navy owned by this faction instead of an agent.
• Fixed blank missions panel being shown in the Prologue.
• When the player choses to withdraw from an enemy attack during the End Turn sequence in Campaign Modes, the "Army Withdraws" message will no longer be received twice.
• In the Prologue Campaign, the Governor slot on the Family Tree panel will no longer overlap the details panel to its right.
• Tweaked the sensitivity of campaign map tooltips, to fix issue where would immediately disappear when opening up due to subtle camera movements.
• Siege attrition is now updated on the Pre-battle panel and tool-tip, so it's now clearer how long a garrison has before supplies run out and they suffer attrition.
• Fixed incorrect display of trade partner values when blockaded.
• The Sea Route icon has been updated so the player can tell it apart from Sea Route Blockaded icon.
• Updated trade breakdown tooltip to have long term relationship value on it.
• Added raided/blockaded to tooltip of resources on the Finance panel, as red state was a bit unclear.
• Added raided/blockaded texts for trade tooltips in Campaign Modes.
• Fixed Total Export label text shrinking prematurely on the Finance Panel in Campaign modes.
• Fixed incorrect positioning of faction icon on agents on the Tactical Map in Campaign modes when no regions owned and no faction leader.
• Fixed spy effect icons spilling out of the Settlement panel in Campaign modes, the icons now overlap when necessary to fit in the available space.
• The "Influence Breakdown" subheading on the Province Overview panel in Campaign Modes now fits better in the available space.
• Added more room for text on the campaign dilemmas user interface.
• Fix for pre-battle camera location loading incorrectly in Legendary difficulty campaign autosaves.
• The target for a War co-ordinator can now be set by clicking on the army/faction icon in Campaign modes.
• The encyclopaedia is now closed when a fullscreen movie is played.

General
• Correct number of SSR options will now be displayed in the game’s settings menus.
• Non-migratory Barbarian City and Towns building descriptions have been updated.
• Added missing cursor variants in the Mac version of the game.
• The opening cinematic at the start of the game will no longer freeze when the player minimises the game window.
• Fix for loading screen sometimes not having any audio.
• Improved advisor lip syncing in Italian.
• Port building chain descriptions have been corrected.
• Fixed the label for the "Show unit banners" (Zobrazovat praporce jednotek) option in the Game Settings in the Czech language version.
Quelle: http://wiki.totalwar.com/w/Update

Edit: Gerade im Total war Forum gelesen, dass die drei neuen DLC-Fraktionen nun auch im spiel integriert sind...
 
Zuletzt bearbeitet:
Zumindest der MSAA fix für AMD Karten scheint zu funktionieren.

Hab jetzt fast 50% mehr Performance auf einer 7970:
von 15fps auf 21fps (extreme settings)

21fps ist zwar nicht spielbar aber ich nehme an, das die Leistung auf den 290ern auch besser ist.

Und wenn jetzt noch das flickern im CF Modus behoben wird, dann kann ich auch extreme settings benutzen.
 
Das Spiel ist immer noch ein Witz. Bei der Gammelgrafik solche FPS Probleme...
 
Hi Leute ich habe meinen Attila Benchmark Thread auf Total War Center upgedated.

Es war uns ja schon immer bekannt wie schlecht AMD CPUs mit Total War funktionieren. Nun aber haben wir zum ersten mal einen Direktvergleich, weil ein User seinen AMD CPU / Mainboard gegen einen Intel CPU getauscht hat. RAM und Grafikkarte sowie die Windows Installation blieben identisch.

Sobald die GPU Limitierung auf niedrigeren Einstellungen verschwunden ist, wird das ganze Desaster sichtbar.


Attila 1.0 Benchmark:

max performance settings:
max 134 / min 58 / avg 86,5 - LordDGP, Intel Core i7-4790K, 4.4 GHz, Asus Maximus VII Hero, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit
max 86 / min 29 / avg 50,8 - LordDGP, AMD Phenom II X6 1100T, 4.0 GHz, Asus Crosshair IV Formula, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit

performance setttings:
max 64 / min 34 / avg 48,0 - LordDGP, Intel Core i7-4790K, 4.4 GHz, Asus Maximus VII Hero, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit
max 46 / min 26 / avg 36,3 - LordDGP, AMD Phenom II X6 1100T, 4.0 GHz, Asus Crosshair IV Formula, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit

quality settings:
max 64 / min 34 / avg 48,0 - LordDGP, Intel Core i7-4790K, 4.4 GHz, Asus Maximus VII Hero, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit
max 36 / min 18 / avg 26,7 - LordDGP, AMD Phenom II X6 1100T, 4.0 GHz, Asus Crosshair IV Formula, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit

maximum quality:
max 48 / min 31 / avg 39,9 - LordDGP, Intel Core i7-4790K, 4.4 GHz, Asus Maximus VII Hero, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit
max 29 / min 14 / avg 22,4 - LordDGP, AMD Phenom II X6 1100T, 4.0 GHz, Asus Crosshair IV Formula, 8 GB RAM 1333 MHz CL 9-9-9-24 1T@1.50 Volts DualChannel, Gigabyte Radeon R9 290 WF 3X OC, Windows 7 Home Premium SP1 64bit


Der angebliche Performance Patch für AMD GPUs Attila 1.1 brachte leider sogar schlechtere Performance bei mehreren PCs (durchschnittlich -2 FPS). Nur MSAA ist laut einem Experten performanter geworden.
 
punkrockfan schrieb:

darum ging es aber in dem Vergleich. ;) Er hat ja nicht seine AMD Grafikkarte getauscht, sondern die CPU.
Achso du redest von dem Wundertreiber wegen des Performance Patches... jetzt klingelt es. Naja ich hab da wenig Hoffnung. Aber gerne kannst du ja mal einen vorher nacher Vergleich machen mit dem Benchmark.
 
Hallo Leute,

es gab bei mir ein günstiges Upgrade und ich bin jetzt fast stolzer Besitzer einer GTX 970 mit der man nun auch Attila spielen könnte.

Für selbiges Spiel, wissen vielleicht einige, betreibe ich schon seit Shogun 2 einen Benchmarkthread in dem jeder willkommen ist und mitbenchen darf. Es geht hier nicht um das setzen von Rekorden, sondern darum zu ersehen wie denn die Leistung auf "Otto-Normal" PCs und vor allem Laptops ist.


Für PCs und auch Laptops inkl. Intel HD Karten haben wir jetzt neue Ergebnisse. Viel Spaß beim Lesen!
http://www.twcenter.net/forums/show...April-2015&p=14360595&viewfull=1#post14360595


Viele Grüße
al Qamar
 
Zurück
Oben