|
|
@@ -836,8 +836,7 @@ def CompileCxx(obj,src,opts):
|
|
|
if (optlevel==3): cmd += " /MD /Zi /O2 /Ob2 /Oi /Ot /fp:fast /DFORCE_INLINING"
|
|
|
if (optlevel==4):
|
|
|
cmd += " /MD /Zi /Ox /Ob2 /Oi /Ot /fp:fast /DFORCE_INLINING /DNDEBUG /GL"
|
|
|
- cmd += " /Oy" # jcr add
|
|
|
- cmd += " /Zp16" # jcr add # Is this necessary with /Ox?
|
|
|
+ cmd += " /Oy /Zp16" # jean-claude add /Zp16 insures correct static alignment for SSEx
|
|
|
|
|
|
cmd += " /Fd" + os.path.splitext(obj)[0] + ".pdb"
|
|
|
building = GetValueOption(opts, "BUILDING:")
|
|
|
@@ -878,17 +877,12 @@ def CompileCxx(obj,src,opts):
|
|
|
if (optlevel==3):
|
|
|
cmd += " /MD /Zi /O2 /Oi /Ot /arch:SSE3"
|
|
|
cmd += " /Ob0"
|
|
|
- cmd += " /Qipo-" # beware of IPO !!!
|
|
|
+ cmd += " /Qipo-" # beware of IPO !!!
|
|
|
## Lesson learned: Don't use /GL flag -> end result is MESSY
|
|
|
## ----------------------------------------------------------------
|
|
|
- ## if (optlevel==4): cmd += " /MD /Zi /O3 /Oi /Ot /arch:SSE3 /Yc /DNDEBUG"
|
|
|
if (optlevel==4):
|
|
|
- cmd += " /MD /Zi /O3 /Oi /Ot /Ob0 /Yc /DNDEBUG" # /Ob0 a ete rajoute en cours de route a 47%
|
|
|
- # cmd += " /MD /Zi /Ox /Oi /Ot /Ob0 /Yc /DNDEBUG" # /Ob0 a ete rajoute en cours de route a 47% test Ox
|
|
|
-
|
|
|
- # cmd += " /Qipo-"
|
|
|
- # cmd += " /Qip" # optimization mono file
|
|
|
- cmd += " /Qipo" # optimization multi file
|
|
|
+ cmd += " /MD /Zi /O3 /Oi /Ot /Ob0 /Yc /DNDEBUG" # /Ob0 a ete rajoute en cours de route a 47%
|
|
|
+ cmd += " /Qipo" # optimization multi file
|
|
|
|
|
|
# for 3 & 4 optimization levels
|
|
|
# -----------------------------
|
|
|
@@ -896,33 +890,32 @@ def CompileCxx(obj,src,opts):
|
|
|
cmd += " /fp:fast=2"
|
|
|
cmd += " /Qftz"
|
|
|
cmd += " /Qfp-speculation:fast"
|
|
|
- cmd += " /Qopt-matmul" # needs /O2
|
|
|
+ cmd += " /Qopt-matmul" # needs /O2 or /O3
|
|
|
cmd += " /Qprec-div-"
|
|
|
cmd += " /Qsimd"
|
|
|
|
|
|
- cmd += " /QxHost" # compile for target host
|
|
|
- cmd += " /Quse-intel-optimized-headers" # use intel optimized headers
|
|
|
- cmd += " /Qparallel" # enable parallelization
|
|
|
- cmd += " /Qvc9" # for Microsoft Visual C++ 2008
|
|
|
-
|
|
|
- # cmd += " /Qopt-report:2 /Qopt-report-phase:hlo /Qopt-report-phase:hpo" # some optimization reports
|
|
|
+ cmd += " /QxHost" # compile for target host; Compiling for distribs should probably strictly enforce /arch:..
|
|
|
+ cmd += " /Quse-intel-optimized-headers" # use intel optimized headers
|
|
|
+ cmd += " /Qparallel" # enable parallelization
|
|
|
+ cmd += " /Qvc9" # for Microsoft Visual C++ 2008
|
|
|
|
|
|
## PCH files coexistence: the /Qpchi option causes the Intel C++ Compiler to name its
|
|
|
## PCH files with a .pchi filename suffix and reduce build time.
|
|
|
## The /Qpchi option is on by default but interferes with Microsoft libs; so use /Qpchi- to turn it off.
|
|
|
- cmd += " /Qpchi-" # keep it this way!
|
|
|
+ ## I need to have a deeper look at this since the compile time is quite influenced by this setting !!!
|
|
|
+ cmd += " /Qpchi-" # keep it this way!
|
|
|
|
|
|
- ## Inlining seems to be an issue here !
|
|
|
+ ## Inlining seems to be an issue here ! (the linker doesn't find necessary info later on)
|
|
|
## ------------------------------------
|
|
|
- ## so don't use cmd += " /DFORCE_INLINING" (need to check why with Panda developpers!)
|
|
|
- ## Inline expansion /Ob1 : Allow functions marked inline to be inline.
|
|
|
- ## Inline any /Ob2 : Inline functions deemed appropriate by compiler.
|
|
|
+ ## so don't use cmd += " /DFORCE_INLINING" (need to check why with Panda developpers!)
|
|
|
+ ## Inline expansion /Ob1 : Allow functions marked inline to be inlined.
|
|
|
+ ## Inline any /Ob2 : Inline functions deemed appropriate by compiler.
|
|
|
|
|
|
- ## Ctor displacement /vd0 : Disable constructor displacement.
|
|
|
+ ## Ctor displacement /vd0 : Disable constructor displacement.
|
|
|
## Choose this option only if no class constructors or destructors call virtual functions.
|
|
|
## Use /vd1 (default) to enable. Alternate: #pragma vtordisp
|
|
|
|
|
|
- ## Best case ptrs /vmb : Use best case "pointer to class member" representation.
|
|
|
+ ## Best case ptrs /vmb : Use best case "pointer to class member" representation.
|
|
|
## Use this option if you always define a class before you declare a pointer to a member of the class.
|
|
|
## The compiler will issue an error if it encounters a pointer declaration before the class is defined.
|
|
|
## Alternate: #pragma pointers_to_members
|
|
|
@@ -933,11 +926,16 @@ def CompileCxx(obj,src,opts):
|
|
|
if ("BIGOBJ" in opts) or (is_64):
|
|
|
cmd += " /bigobj"
|
|
|
|
|
|
- # level of warnings
|
|
|
- cmd += " /EHa /Zm300 /DWIN32_VC /DWIN32 /W3 " + BracketNameWithQuotes(src)
|
|
|
- # cmd += " /EHa /Zm300 /DWIN32_VC /DWIN32 /W4 " + BracketNameWithQuotes(src)
|
|
|
- # cmd += " /EHa /Zm300 /DWIN32_VC /DWIN32 /Wall " + BracketNameWithQuotes(src)
|
|
|
+ # level of warnings and optimization reports
|
|
|
+ if GetVerbose():
|
|
|
+ cmd += " /W3 " # or /W4 or /Wall
|
|
|
+ cmd += " /Qopt-report:2 /Qopt-report-phase:hlo /Qopt-report-phase:hpo" # some optimization reports
|
|
|
+ else:
|
|
|
+ cmd += " /W1 "
|
|
|
+ cmd += " /EHa /Zm300 /DWIN32_VC /DWIN32 " + BracketNameWithQuotes(src)
|
|
|
+
|
|
|
oscmd(cmd)
|
|
|
+
|
|
|
if (COMPILER=="LINUX"):
|
|
|
cc = os.environ.get('CC', 'gcc')
|
|
|
cxx = os.environ.get('CXX', 'g++')
|
|
|
@@ -1183,7 +1181,8 @@ def CompileLink(dll, obj, opts):
|
|
|
else: mtcmd = mtcmd + ";1"
|
|
|
oscmd(mtcmd, ignoreError=True) # HACK: For some reason, mt sometimes gives a non-zero return value, even when it works
|
|
|
else:
|
|
|
- cmd = "xilink /verbose:lib"
|
|
|
+ cmd = "xilink"
|
|
|
+ if GetVerbose(): cmd += " /verbose:lib"
|
|
|
if (is_64):
|
|
|
cmd += " /MACHINE:X64"
|
|
|
if ("MFC" not in opts):
|
Zhao Huang