Changes between Version 8 and Version 9 of GcnInstrsVop1


Ignore:
Timestamp:
Nov 29, 2015, 12:00:17 AM (4 years ago)
Author:
trac
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • GcnInstrsVop1

    v8 v9  
    13501350Syntax: V_RCP_CLAMP_F32 VDST, SRC0<br />
    13511351Description: Approximate reciprocal from floating point value SRC0 and store it to VDST.
    1352 Guaranted error below 1ulp. Result is clamped maximum float value including its sign.<br />
     1352Guaranted error below 1ulp. Result is clamped to MAX_FLOAT including sign of a result.<br />
    13531353Description:<br />
    13541354<code>VDST = APPROX_RCP(ASFLOAT(SRC0))
     
    13631363Description:<br />
    13641364<code>VDST = APPROX_RCP(ASFLOAT(SRC0))</code></p>
     1365<h4>V_RCP_F64</h4>
     1366<p>Opcode VOP1: 47 (0x2f) for GCN 1.0/1.1; 37 (0x25) for GCN 2.0<br />
     1367Opcode VOP3A: 431 (0x1af) for GCN 1.0/1.1; 357 (0x165) for GCN 2.0<br />
     1368Syntax: V_RCP_F64 VDST(2), SRC0(2)<br />
     1369Description: Approximate reciprocal from double FP value SRC0 and store it to VDST.
     1370Relative error of approximation is ~1e-8.<br />
     1371Description:<br />
     1372<code>VDST = APPROX_RCP(ASDOUBLE(SRC0))</code></p>
     1373<h4>V_RCP_CLAMP_F64</h4>
     1374<p>Opcode VOP1: 48 (0x30) for GCN 1.0/1.1<br />
     1375Opcode VOP3A: 432 (0x1b0) for GCN 1.0/1.1<br />
     1376Syntax: V_RCP_CLAMP_F64 VDST(2), SRC0(2)<br />
     1377Description: Approximate reciprocal from double FP value SRC0 and store it to VDST.
     1378Relative error of approximation is ~1e-8.
     1379Result is clamped to MAX_DOUBLE value including sign of a result.<br />
     1380Description:<br />
     1381<code>VDST = APPROX_RCP(ASDOUBLE(SRC0))
     1382if (ABS(ASDOUBLE(VDST))==INF)
     1383    VDST = SIGN(ASDOUBLE(VDST)) * MAX_DOUBLE</code></p>
    13651384<h4>V_RCP_IFLAG_F32</h4>
    13661385<p>Opcode VOP1: 43 (0x2b) for GCN 1.0/1.1; 35 (0x23) for GCN 2.0<br />
     
    14211440if (ASFLOAT(VDST)==INF)
    14221441    VDST = MAX_FLOAT</code></p>
     1442<h4>V_RSQ_F32</h4>
     1443<p>Opcode VOP1: 46 (0x2e) for GCN 1.0/1.1; 36 (0x24) for GCN 2.0<br />
     1444Opcode VOP3A: 430 (0x1ae) for GCN 1.0/1.1; 356 (0x164) for GCN 2.0<br />
     1445Syntax: V_RCP_F32 VDST, SRC0<br />
     1446Description: Approximate reciprocal square root from floating point value SRC0 and
     1447store it to VDST. If SRC0 is negative value, store -NAN to VDST.
     1448This instruction doesn't handle denormalized values regardless FLOAT MODE register setup.<br />
     1449Description:<br />
     1450<code>VDST = APPROX_RSQRT(ASFLOAT(SRC0))</code></p>
    14231451<h4>V_RSQ_LEGACY_F32</h4>
    14241452<p>Opcode VOP1: 45 (0x2d) for GCN 1.0/1.1<br />
     
    14331461if (ASFLOAT(VDST)==INF)
    14341462    VDST = 0.0</code></p>
    1435 <h4>V_RSQ_F32</h4>
    1436 <p>Opcode VOP1: 46 (0x2e) for GCN 1.0/1.1; 36 (0x24) for GCN 2.0<br />
    1437 Opcode VOP3A: 430 (0x1ae) for GCN 1.0/1.1; 356 (0x164) for GCN 2.0<br />
    1438 Syntax: V_RCP_F32 VDST, SRC0<br />
    1439 Description: Approximate reciprocal square root from floating point value SRC0 and
    1440 store it to VDST. If SRC0 is negative value, store -NAN to VDST.
    1441 This instruction doesn't handle denormalized values regardless FLOAT MODE register setup.<br />
    1442 Description:<br />
    1443 <code>VDST = APPROX_RSQRT(ASFLOAT(SRC0))</code></p>
    14441463<h4>V_TRUNC_F32</h4>
    14451464<p>Opcode VOP1: 33 (0x21) for GCN 1.0/1.1; 28 (0x1c) for GCN 1.2<br />